Group recommender systems for tourism: overcoming the cold-start problem and conflicting preferences by using artificial intelligence and serious games

Uni e sidade do Minho
Escola de Engenha ia
Pa ícia Raquel de Jesus A aújo Al es
G oup Recommende Sys ems o Tou ism:
O e coming he Cold-S a P oblem and
Con lic ing P e e ences by using A i icial
In elligence and Se ious Games
Decembe 2024
UMinho | 2024 Pa ícia Raquel de Jesus A aújo Al es
G oup Recommende Sys ems o Tou ism: O e coming he
Cold-S a P oblem and Con lic ing P e e ences by using
A i icial In elligence and Se ious Games
Uni e sidade do Minho
Escola de Engenha ia
Pa ícia Raquel de Jesus A aújo Al es
G oup Recommende Sys ems o
Tou ism: O e coming he Cold-
S a P oblem and Con lic ing
P e e ences by using A i icial
In elligence and Se ious Games
Doc o al Thesis
Doc o al P og am in In o ma ics
Wo k pe o med unde he supe ision o :
P o esso Paulo No ais
P o esso a Go e i Ma ei os
Decembe 2024
ii
COPYRIGHT AND CONDITIONS OF USE OF THE WORK BY THIRD PARTIES
This is an academic wo k ha may be used by hi d pa ies p o ided ha in e na ionally accep ed
ules and good p ac ices ega ding copy igh and ela ed igh s a e espec ed.
This wo k may he e o e be used in acco dance wi h he e ms o he license below.
I he use needs pe mission o use he wo k unde condi ions no p o ided o in he license indi-
ca ed, hey should con ac he au ho h ough he Uni e si y o Minho Reposi o iUM.
License g an ed o use s o his wo k
1
A ibu ion-NonComme cial-NoDe i a i es 4.0 In e na ional
CC BY-NC-ND 4.0
h ps://c ea i ecommons.o g/licenses/by-nc-nd/4.0/
1
Based on he mos es ic i e license o he published pape s.
iii
To Júnio and Ma alda, I go his a because o you!
To he amazing people who ca ied me o he ligh a he end o he unnel.

i
STATEMENT OF INTEGRITY
I he eby decla e ha ing conduc ed his academic wo k wi h in eg i y. I con i m ha I ha e no used
plagia ism o any o m o undue use o in o ma ion o alsi ica ion o esul s along he p ocess
leading o i s elabo a ion.
I u he decla e ha I ha e ully acknowledged he Code o E hical Conduc o he Uni e si y o
Minho.
Po o, Decembe 22, 2024
__________________________________________________
(Pa ícia Raquel de Jesus A aújo Al es)
A C K N O W L E D G E M E N T S
Al hough i doesn’ seem, I’m no always keen on ou ines and he o dina y, and I’m known o
being pe se e an and “s ubbo n”. Some imes I like o expe ience di e en pa hs, aim o he
highes s akes, and do hings di e en ly om he common. So, as hey a e pe sonal, I decided o
gi e a pe sonal ouch in he acknowledgemen s o my “pe sonal agen s”.
i
This wo k was suppo ed by he Eu opean Social Fund and FCT unde G an 2020.06129.BD
ii
P R E F A C E
I ne e eally liked memo izing ma e ial o es s, al hough ha ing a e y good memo y (well, now
i ’s no ha good... oo much sleep dep i a ions and insomnias…:’)). Bu I ha e always enjoyed
eading and esea ching, and I lo e science, he elemen a y, he beginning o hings. Al hough o
s udy Medicine was in he op o my lis in my adolescence, as I ha e always wan ed o be a Su -
geon and help o he s, I ealized I wouldn’ ha e he s omach o some hings. Tha is why I chose
o s udy Chemis y a FCUP in 1998. I jus lo ed o abso b ha knowledge, he heo y o e e y-
hing, o be in he lab, making expe imen s and w i ing epo s on he esul s.
I am so passiona e o compu e s ha , du ing he Chemis y cou se, I had al eady p og ammed in
Pascal my own MS-DOS applica ion o con olling my inances, and a he cou se’s second yea , I
was hinking o changing o he In o ma ics Enginee ing cou se a FEUP, bu I was enjoying Chem-
is y so much... So, I said o mysel I had o inish wha I s a ed! My mo he has always old me o
ne e gi e up and always do my bes , so I con inued in Chemis y un il he end. And I ne e eg e
ha decision, I lea ned so much and me spec acula p o esso s, iends and s a !
I was he so-called compu e geek (I s ill am), always pa icipa ing in LAN pa ies and game e en s
wi h my iends, playing Un eal Tou namen 99, and o he i s -pe son shoo e s like Coun e s ike,
and Quake. And o cou se, all he Tomb Raide and Mys se ies :)
I ha e always been eage o lea n mo e abou di e en a eas. So, I also comple ed a echnological
specializa ion cou se in Mic obiology in he mean ime. A e g adua ing in Chemis y in 2003, I
was planning o do a mas e ’s in As onomy and/o A cheology. I am an As onomy en husias ! Oh
A cheology… Howe e , my oldes child was bo n soon a e , and when I hough I could p oceed
wi h my “knowledge sp ee”, my mo he said, “young lady, now you ha e o p o ide o you sel .”
And so, I did, I wo ked in se e al places, including wo yea s a Glin HS, a heal hca e so wa e
company. Howe e , I el he need o know mo e abou p og amming and in o ma ics. So, I qui
my job o pu sue ha objec i e. F om 2009-2014 I a ended he In eg a ed Mas e ’s in In o ma ics
and Compu ing Enginee ing a FEUP. I eally enjoyed he cou se and lea ned a lo , ha ing o ced
mysel o do some wo ks alone so I could lea n mo e.
Con en s
xi
9.4 RESULTS AND ANALYSIS ..................................................................................................... 251
9.4.1
Pa icipan s Cha ac e iza ion .................................................................................. 251
9.4.2
Expe imen s Resul s and Discussion ....................................................................... 254
9.5 CONCLUSIONS AND FUTURE WORK ....................................................................................... 266
PART III: IN CONCLUSION ........................................................................................ 270
10 CONCLUSIONS ........................................................................................................ 272
10.1 LIMITATIONS ................................................................................................................. 276
10.2 FUTURE DIRECTIONS ...................................................................................................... 277
REFERENCES ............................................................................................................... 280
APPENDIX A: MIND MAP FOR THE MOST RELEVANT TERMS FOR TOURIST ATTRACTIONS ................ 305
APPENDIX B: THE BIG FIVE INVENTORY ............................................................................. 308
APPENDIX C: POSTER PRESENTED AT RECSYS’23 ............................................................... 309
APPENDIX D: PAPER 5 APPENDICES ................................................................................. 310
APPENDIX D-A. CLUSTERING STATISTICS ......................................................................................... 310
APPENDIX D-B. PARTICIPANTS’ REAL RATINGS .................................................................................. 312
APPENDIX D-C. TOURIST ATTRACTION TYPES .................................................................................... 322
APPENDIX D-D. SUPPLEMENTARY DATA........................................................................................... 324
APPENDIX E: PAPER 7 APPENDICES ................................................................................. 325
APPENDIX E-A: WHICH WAY GAME’S RELEVANT STATISTICS ................................................................. 325
APPENDIX E-B: TIME TRAVEL MANIA GAME’S RELEVANT STATISTICS (COMPLETE ATTEMPT) ......................... 329

x
L I S T O F F I G U R E S
FIGURE 1.1. DIAGRAM OF THE DESIGN SCIENCE RESEARCH METHODOLOGY. RETRIEVED WITH PERMISSION FROM
PEFFERS, TUUNANEN, ROTHENBERGER, AND CHATTERJEE (2007). ........................................................... 8
FIGURE 2.1. (LEFT) EXAMPLE OF MUSIC FROM A SPOTIFY RECOMMENDATION LIST BASED ON A PERSONAL PLAYLIST;
(RIGHT) PART OF THE STEREOTYPE GRAPH USED BY GRUNDY RS, ADAPTED FROM RICH (1979). .................... 19
FIGURE 3.1. (LEFT) ARCHITECTURE OF THE CONCEPTUAL MOBILE GROUP RECOMMENDER SYSTEM. (RIGHT)
INFORMATION ABOUT THE TOURISTS, AVAILABLE IN THE MULTI-AGENT SERVICE. .......................................... 54
FIGURE 4.1. CONFIRMATORY FACTOR ANALYSIS OF THE TOURIST ATTRACTIONS PREFERENCE MODEL OBTAINED
WITH THE EXPLORATORY FACTOR ANALYSIS PROCEDURES, SHOWING THE NORMALIZED REGRESSION WEIGHTS
FOR EACH ITEM ............................................................................................................................. 72
FIGURE 4.2. DISTRIBUTION OF THE FIVE PERSONALITY DIMENSIONS SCORES AMONG THE SAMPLE........................... 73
FIGURE 4.3. SIMPLIFIED STRUCTURAL EQUATION MODEL FOR THE PROPOSED “PERSONALITY-TOURIST
ATTRACTIONS PREFERENCE” MODEL. ................................................................................................. 74
FIGURE 5.1. RESPONDENTS’ (A) AGE RANGE, (B) FORMATION AREAS, (C) LIQUID INCOME, (D) LEISURE TRAVELS
ABROAD PER YEAR; 𝑛=1035 ......................................................................................................... 105
FIGURE 5.2. DISTRIBUTION OF THE FIVE PERSONALITY DIMENSIONS RESPONSES (PARTICIPANTS’ MEAN VALUE) ........ 106
FIGURE 5.3. SIMPLIFIED STRUCTURAL EQUATION MODEL FOR THE PROPOSED “TOURISM CATEGORIES” MODEL,
OBTAINED USING CFA .................................................................................................................. 118
FIGURE 5.4. SIMPLIFIED STRUCTURAL EQUATION MODEL FOR THE PROPOSED “PERSONALITY VS TOURIST
ATTRACTIONS PREFERENCE” MODEL, OBTAINED USING CFA ................................................................. 120
FIGURE 5.5. SIMPLIFIED STRUCTURAL EQUATION MODEL FOR THE PROPOSED “TRAVELLING MOTIVATIONS” MODEL,
OBTAINED USING CFA .................................................................................................................. 128
FIGURE 5.6. SIMPLIFIED STRUCTURAL EQUATION MODEL FOR THE PROPOSED “PERSONALITY VS TRAVELLING
MOTIVATIONS” MODEL, OBTAINED USING CFA ................................................................................... 129
FIGURE 5.7. SIMPLIFIED STRUCTURAL EQUATION MODEL FOR THE PROPOSED “TRAVEL-RELATED PREFERENCES AND
CONCERNS” MODEL, OBTAINED USING CFA. THE ITEMS LABEL CAN BE FOUND AT TABLE 5.11 .................... 131
FIGURE 5.8. SIMPLIFIED STRUCTURAL EQUATION MODEL FOR THE PROPOSED “PERSONALITY VS TRAVEL-RELATED
PREFERENCES AND CONCERNS” MODEL, OBTAINED USING CFA ............................................................ 132
FIGURE 6.1. GROUPLANNER MICROSERVICES ARCHITECTURE. ...................................................................... 147
FIGURE 6.2. (LEFT) GROUPLANNER USER INTERFACE SHOWING THE 3 SUBGROUPS FORMED; CENTER: DETAIL OF
THE MEMBERS IN THE “ADRENALINE ACTIVITIES” SUBGROUP FORMED. (RIGHT) LIST OF POI RECOMMENDED
TO THE “ADRENALINE ACTIVITIES” SUBGROUP. .................................................................................. 149
FIGURE 7.1. GROUPLANNER MICROSERVICES ARCHITECTURE (FIRST PUBLISHED IN ALVES ET AL. (2022)). ........... 166
FIGURE 7.2. SOME GROUPLANNER APP SCREENS: (A) NEW USER REGISTRATION FORM; (B) PERSONALITY INVENTORY
(BFI); (C) THE TOURIST’S PERSONALITY SCORES AFTER FILLING THE BFI; (D) THE TOURIST’S PREDICTED
ATTRACTION PREFERENCE; AND (E) THE TOURIST’S PREDICTED TRAVEL-RELATED PREFERENCES & CONCERNS. .. 169
FIGURE 7.3. REPRESENTATION OF THE MAIN EXCURSION DIVISION INTO SUBGROUPS WITH AT LEAST 3 TOURISTS
Lis o Figu es
x i
FROM THE SAME CLUSTER. ............................................................................................................ 176
FIGURE 7.4. DIAGRAM REPRESENTING THE REQUEST OF A GROUP RECOMMENDATION TO THE MAMS. .................. 178
FIGURE 7.5. DIAGRAM REPRESENTING THE CONTINUATION OF THE GROUP RECOMMENDATION REQUEST. ............... 179
FIGURE 7.6. SOME GROUPLANNER’S APP SCREENS USED IN THE RECOMMENDATION PROCESS. ........................... 185
FIGURE 7.7. (A) TIME TAKEN, IN SECONDS, BY EACH ALGORITHM TO CLUSTER 35 TO 100K USERS, FOR 𝑠𝑖𝑚≥
0.80; (B) TIME TAKEN, IN SECONDS, BY EACH ALGORITHM TO ASSIGN A NEW USER WHEN THERE WERE
ALREADY 𝑘 CLUSTERS. ................................................................................................................. 191
FIGURE 7.8. (A) TIME TAKEN, IN SECONDS, BY 𝑑-MEANS TO CLUSTER 35 TO 100K USERS FOR DIFFERENT
SENSITIVITIES; (B) NUMBER OF CLUSTERS FORMED BY 𝑑-MEANS FOR AN INCREASING NUMBER OF USERS IN THE
DATABASE FOR EACH SENSITIVITY. ................................................................................................... 192
FIGURE 7.9. (A) BOXPLOTS COMPARING THE 𝑑-MEANS CLUSTERS COMPACTNESS FOR THE SIMILARITY THRESHOLDS
OF 𝑠𝑖𝑚≥ 0.70 AND 𝑠𝑖𝑚≥ 0.80 FOR 50000 USERS. (B) BOXPLOTS COMPARING THE 𝑑-MEANS
CLUSTERS SILHOUETTE FOR THE SIMILARITY THRESHOLDS OF 𝑠𝑖𝑚≥ 0.70 AND 𝑠𝑖𝑚≥ 0.80 FOR 50000
USERS....................................................................................................................................... 193
FIGURE 7.10. (A) BOXPLOTS COMPARING THE 3 ALGORITHMS IN TERMS OF THE CLUSTERS’ USERS’ AVERAGE
DISTANCE TO THE RESPECTIVE CENTROID; (B) CLOSEST CLUSTER; (C) SILHOUETTE VALUES, FROM 35 TO 100K
USERS, FOR 𝑠𝑖𝑚≥ 0.80; AND (D) 𝑑-MEANS NUMBER OF USERS PER CLUSTER FOR EACH SENSITIVITY, FROM
1000 TO 100K USERS................................................................................................................. 195
FIGURE 7.11. COMPARISON OF THE CLUSTERS 2D SPATIAL DISTRIBUTION FOR 𝑑-MEANS AND THE 2 BASELINES,
FOR N= 50 AND N = 150. ............................................................................................................ 196
FIGURE 7.12. PARTICIPANTS’ (A) AGE RANGE, (B) FEARS/PHOBIAS; 𝑛 = 35. .................................................. 197
FIGURE 7.13. DISTRIBUTION OF THE PARTICIPANTS’ FIVE PERSONALITY DIMENSIONS SCORES. ............................. 200
FIGURE 7.14. (A) PARTICIPANTS’ CLUSTERS’ AVERAGE Α; (B) PARTICIPANTS’ CLUSTERS’ SILHOUETTE; (C)
PARTICIPANTS’ EUCLIDEAN SIMILARITY TO THE CORRESPONDING CLUSTER’S CENTROID, GROUPED BY THEIR
RESPECTIVE CLUSTER. .................................................................................................................. 201
FIGURE 7.15. (A) FORMED SUBGROUPS AVERAGE DISTANCE TO THE CORRESPONDING SUBGROUP’S CENTROID; (B)
FORMED SUBGROUPS SILHOUETTE VALUES. ...................................................................................... 208
FIGURE 7.16. EXAMPLE OF SOME RULES FOUND WITH THE APRIORI ALGORITHM. .............................................. 208
FIGURE 7.17. SUBGROUPS FORMED DISPLAYED AS A RADAR CHART OF THE CORRESPONDING MEMBERS’
PERSONALITY. ............................................................................................................................. 209
FIGURE 7.18. (A) RATINGS (1 TO 5 STARS) GIVEN BY THE PARTICIPANTS TO THE INDIVIDUAL POI
RECOMMENDATIONS; (B) AVERAGE RATINGS GIVEN TO THE SUGGESTED POI BY THE EXCURSION SUBGROUPS’
MEMBERS. ................................................................................................................................. 214
FIGURE 8.1. REGISTERING A TOURIST USER IN THE APPLICATION WITH DYNAMIC CLUSTERING (SIMPLIFIED DIAGRAM). 226
FIGURE 8.2. SIMPLIFIED DIAGRAM FOR THE GROUP RECOMMENDATION PROCESS. ............................................. 227
FIGURE 9.1. (LEFT) INITIAL SCENE OF THE CAUTIOUSNESS PERSONALITY MINIGAME. (RIGHT) ACHIEVEMENT-STRIVING
MINIGAME INITIAL SCREEN. ............................................................................................................ 242
FIGURE 9.2. CAUTIOUSNESS GAME SCENES. (LEFT) ONE OF THE HIDDEN MS. (RIGHT) EXAMPLE OF A BIFURCATION.
THE TOP PATH IS AN EASY PATH, AND THE BOTTOM ONE, A HARD PATH, WITH A HEALTH POTION AT THE END.
THE VERTICAL RED LINES MARK THE INVISIBLE WALLS. ......................................................................... 243
Lis o Figu es
x ii
FIGURE 9.3. EXAMPLE OF A TRAP WITH POISONOUS GAS ACTIVATED (PINK FOG) IN THE CAUTIOUSNESS MINIGAME. ... 244
FIGURE 9.4. SCENE OF A HARD PATH SHOWING FOG AND AN AI MOVED ENEMY IN THE CAUTIOUSNESS GAME. ......... 244
FIGURE 9.5. (LEFT) ACHIEVEMENT-STRIVING GAME START MENU. (RIGHT) ACHIEVEMENT-STRIVING GAME END
PANEL. ...................................................................................................................................... 245
FIGURE 9.6. SCENES OF THE TIME TRAVEL GAME. (TOP LEFT) A LONG AND TEDIOUS PATH THAT ALLOWS CATCHING
MANY HARD COINS AND A DIAMOND, ONLY PARTIALLY SEEN AS THE PLAYER ADVANCES. (TOP RIGHT-TOP) PATH
HIDDEN IN THE CLOUDS LEADING TO A DIAMOND. (TOP RIGHT-BOTTOM) BIFURCATION WITH AN EASY AND HARD
PATH. (BOTTOM) PART OF A HIDDEN LONG PATH LEADING TO A DIAMOND. ................................................ 246
FIGURE 9.7. TWO PARTICIPANTS PLAYING THE PERSONALITY MINIGAMES. ........................................................ 248
FIGURE 9.8. PARTICIPANTS’ MEAN SCORES FOR SOME PERSONALITY TRAITS IN THE IPIP-NEO-120 QUESTIONNAIRE
(𝑛 = 100). ............................................................................................................................... 253
FIGURE 9.9. HISTOGRAMS FOR THE METRICS MEASURED IN THE WHICH WAY MINIGAME (𝑛=86). ........................ 256
FIGURE 9.10. INDEPENDENT-SAMPLES TESTS FOR THE CAUTIOUSNESS DISTRIBUTION ACROSS DIFFERENT GROUPS
OF THE WW METRICS (𝑛=86). ...................................................................................................... 258
FIGURE 9.11. HISTOGRAMS FOR METRICS MEASURED IN THE TIME TRAVEL MINIGAME (𝑛=70) IN THE FIRST
COMPLETE ATTEMPT. .................................................................................................................... 261
FIGURE 10.1. DIAGRAM REPRESENTING THE RELATIONSHIP BETWEEN THE PROPOSED OBJECTIVES, RESEARCH
QUESTIONS, AND THE SEVEN PAPERS WRITTEN. .................................................................................. 276
x iii
L I S T O F T A B L E S
TABLE 2.1. AGGREGATION STRATEGIES USED IN GRS, ADAPTED FROM MASTHOFF (2015). ................................. 32
TABLE 2.2. COMPARISON OF THE MICROSERVICE PRINCIPLES TO MAS. ADAPTED FROM W. COLLIER ET AL. (2019). . 34
TABLE 2.3. COMPARISON OF THE MAS PRINCIPLES TO MICROSERVICES. ADAPTED FROM W. COLLIER ET AL.
(2019)....................................................................................................................................... 35
TABLE 2.4. COSTA AND MACCRAE (1992) NEO PI-R PERSONALITY DIMENSIONS’ FACETS, ADAPTED FROM JOHN
AND SRIVASTAVA (1999). ............................................................................................................... 37
TABLE 4.1. PERSONALITY DIMENSIONS AND THEIR RESPECTIVE SIX TRAITS (ADAPTED FROM COSTA AND MACCRAE
(1992)). .................................................................................................................................... 64
TABLE 4.2. SAMPLE DESCRIPTIVE STATISTICS (N=508). ............................................................................... 71
TABLE 4.3. VARIMAX ROTATED COMPONENT MATRIX FOR THE PROPOSED TOURIST ATTRACTIONS, SHOWING THE 11
FACTORS EXTRACTED USING EFA, THE ESTIMATED CORRELATIONS BETWEEN THE ITEMS AND FACTORS, AND
EACH FACTORS CRONBACH’S ALPHA RELIABILITY. ................................................................................. 75
TABLE 4.4. PERSONALITY DIMENSIONS MEAN TOTAL SCORES (BFI-44), N=508. ............................................... 76
TABLE 5.1. PERSONALITY DIMENSIONS AND THEIR RESPECTIVE SIX TRAITS (ADAPTED FROM COSTA AND MACCRAE
(1992)) ..................................................................................................................................... 84
TABLE 5.2. SAMPLE DESCRIPTIVE STATISTICS (N=1035) ............................................................................ 104
TABLE 5.3. PARTICIPANTS' PREFERENCES FOR TOURIST ATTRACTIONS, IN PERCENTAGE OF AGREEMENT. ................ 107
TABLE 5.4. PARTICIPANTS' TRAVELLING MOTIVATIONS, IN PERCENTAGE OF AGREEMENT (QUESTIONS ADAPTED FROM
PEARCE AND LEE (2005)). ........................................................................................................... 111
TABLE 5.5. PARTICIPANTS' TRAVEL-RELATED PREFERENCES AND CONCERNS, IN PERCENTAGE OF AGREEMENT. ........ 112
TABLE 5.6. CONFIRMATORY FACTOR ANALYSIS OF THE BIG FIVE INVENTORY RESPONSES, CONFIRMING THE 5
PERSONALITY FACTORS EXTRACTED USING EFA AND THEIR RESPECTIVE ITEMS, THE STANDARDIZED REGRESSION
WEIGHTS BETWEEN THE ITEMS AND FACTORS, AND EACH FACTOR’S CRONBACH’S ALPHA RELIABILITY. ............ 116
TABLE 5.7. VARIMAX ROTATED COMPONENT MATRIX FOR THE PROPOSED TOURISM CATEGORIES, SHOWING THE 11
FACTORS EXTRACTED USING EFA AND THEIR RESPECTIVE ITEMS, THE ESTIMATED CORRELATIONS BETWEEN THE
ITEMS AND FACTORS, AND EACH FACTOR’S CRONBACH’S ALPHA RELIABILITY. ............................................ 116
TABLE 5.8. STANDARDIZED REGRESSION WEIGHTS FOR THE RELATIONSHIP BETWEEN THE BFI DIMENSIONS AND THE
PREFERENCE FOR TOURIST ATTRACTIONS, OBTAINED USING CFA. .......................................................... 121
TABLE 5.9. VARIMAX ROTATED COMPONENT MATRIX FOR THE PROPOSED TRAVELLING MOTIVATIONS, SHOWING THE
6 FACTORS EXTRACTED USING EFA AND THEIR RESPECTIVE ITEMS, THE ESTIMATED CORRELATIONS BETWEEN
THE ITEMS AND FACTORS, AND EACH FACTOR’S CRONBACH’S ALPHA RELIABILITY. ...................................... 125
TABLE 5.10. STANDARDIZED REGRESSION WEIGHTS FOR THE RELATIONSHIP BETWEEN THE BFI DIMENSIONS AND
TRAVELLING MOTIVATIONS, OBTAINED USING CFA. ............................................................................. 128
TABLE 5.11. VARIMAX ROTATED COMPONENT MATRIX FOR THE PROPOSED TRAVEL-RELATED PREFERENCES AND
CONCERNS, SHOWING THE 4 FACTORS EXTRACTED USING EFA AND THEIR RESPECTIVE ITEMS, THE ESTIMATED
CORRELATIONS BETWEEN THE ITEMS AND FACTORS, AND EACH FACTOR’S CRONBACH’S ALPHA RELIABILITY. .... 130
Lis o Tables
xix
TABLE 5.12
.
STANDARDIZED REGRESSION WEIGHTS FOR THE RELATIONSHIP BETWEEN THE BFI DIMENSIONS AND
TRAVEL-RELATED PREFERENCES AND CONCERNS, OBTAINED USING CFA.................................................. 132
TABLE 5.13. CORRELATIONS BETWEEN PERSONALITY, FEARS, AND CLIMATE CONDITIONS PREFERENCE AT THE
DESTINATION. ............................................................................................................................. 133
TABLE 5.14. 𝑑𝑖𝑓𝑆𝑐𝑜𝑟𝑒𝑠𝑇𝑖𝑇𝑗 WEIGHT LEVELS ..................................................................................... 137
TABLE 5.15. POSSIBLE DEGREES OF COMPATIBILITY BETWEEN TOURISTS ....................................................... 137
TABLE 7.1. DESCRIPTION OF THE TOURISM CATEGORIES AVAILABLE FOR RECOMMENDATION ................................ 170
TABLE 7.2. DESCRIPTION OF THE TRAVEL-RELATED PREFERENCES & CONCERNS CONSIDERED IN THE
RECOMMENDATIONS ..................................................................................................................... 171
TABLE 7.3. USERS’ ATTRIBUTES, AND RESPECTIVE CLASSES, USED BY THE MAMS TO DETERMINE THE ASSOCIATION
RULES LIST FOR EACH CLUSTER. ..................................................................................................... 175
TABLE 7.4. POINTS OF INTEREST MAIN ATTRIBUTES. .................................................................................. 182
TABLE 7.5. AVERAGE TIME TAKEN, IN SECONDS, AND NUMBER OF CLUSTERS FORMED BY 𝑑-MEANS TO CLUSTER 35
TO 100K USERS FOR DIFFERENT SENSITIVITIES. .................................................................................. 193
TABLE 7.6. PARTICIPANTS DESCRIPTIVE STATISTICS (N=35). ....................................................................... 199
TABLE 7.7. PERSONALITY, PREDICTED TOURIST ATTRACTION PREFERENCE AND PREDICTED TRAVEL-RELATED
PREFERENCES & CONCERNS OF THE PARTICIPANTS IN THE CLUSTERS FORMED, GROUPED BY THE
CORRESPONDING SUBGROUP (𝐶𝑞𝑆𝑔). ............................................................................................ 203
TABLE 7.8. CLUSTERS FORMED BY THE PROPOSED 𝑑-MEANS ALGORITHM AND THEIR RESPECTIVE AVERAGE DISTANCE
TO THE CENTROID (Α), CLOSEST CLUSTER (Β, SEPARABILITY), SILHOUETTE VALUES (𝑠), AND EUCLIDEAN
SIMILARITY (𝑠𝑖𝑚) BETWEEN THE PARTICIPANT 𝑝𝑖 AND THE CORRESPONDING CLUSTER’S CENTROID. ............. 205
TABLE 7.9. LIST OF POI RECOMMENDED TO EACH SUBGROUP WITH THE RESPECTIVE POI TOURISM CATEGORY, POI
ID, POI DESCRIPTION AND AVERAGE RATING GIVEN BY THE SIMULATION PARTICIPANTS (FROM 1 TO 5 STARS).... 205
TABLE 7.10. LEFT SIDE: PREDICTED AVERAGE PREFERENCE FOR THE TOURISM CATEGORIES AND THE TRAVEL-
RELATED PREFERENCES & CONCERNS FOR THE FORMED SUBGROUPS (𝐶𝑞𝑆𝑔), FROM 0.0 TO 1.0, FOLLOWED
BY THE CORRESPONDING RATING FROM 1 TO 5 STARS IN PARENTHESIS. RIGHT SIDE: AVERAGE RATINGS FOR THE
TOURISM CATEGORIES AND TRAVEL-RELATED PREFERENCES & CONCERNS GIVEN BY THE MEMBERS IN THE PRE-
QUESTIONNAIRE, FROM 1 TO 5 STARS. ............................................................................................. 211
TABLE 7.11. PAIRED SAMPLES T-TEST OF THE MEANS BETWEEN THE REAL RATINGS GIVEN BY THE PARTICIPANTS TO
THE TOURISM CATEGORIES AND THE TRAVEL-RELATED PREFERENCES & CONCERNS IN THE PRE-QUESTIONNAIRE
AND THE PREDICTED ONES. ............................................................................................................ 213
TABLE 9.1. BIG FIVE PERSONALITY DIMENSIONS AND CORRESPONDING SIX TRAITS, ADAPTED FROM COSTA AND
MACCRAE (1992)
.
..................................................................................................................... 235
TABLE 9.2. PARTICIPANTS MOST RELEVANT DESCRIPTIVE STATISTICS (𝑛=100). ............................................... 252
TABLE 9.3. TIME TAKEN BY PARTICIPANTS, IN MINUTES, TO COMPLETE THE GAMES WHEN HELPED AND NOT HELPED. 255
TABLE 9.4. CORRELATION COEFFICIENTS BETWEEN THE WW GAME MEASURED VARIABLES AND RELEVANT
PERSONALITY TRAITS. ................................................................................................................... 257
TABLE 9.5. CORRELATION COEFFICIENTS BETWEEN THE TT GAME MEASURED VARIABLES AND RELEVANT
PERSONALITY TRAITS .................................................................................................................... 263

xx
A B B R E V I A T I O N S A N D A C R O N Y M S
AI
A i icial In elligence
ALGORITMI
Resea ch uni o he School o Enginee ing – Uni e si y o Minho, ha de elops
R&D ac i i y in In o ma ion and Communica ions Technology and Elec onics.
API
Applica ion P og amming In e ace
AR
Augmen ed Reali y
ASOC
Applied So Compu ing
BFI
Big Fi e In en o y
CEOS.PP
Cen e o O ganiza ional and Social S udies o he Poly echnic Ins i u e o Po o
CFA
Con i ma o y Fac o Analysis
CORE
COmpu ing Resea ch & Educa ion
DB
Da abase
DSRM
Design Science Resea ch Me hodology
EFA
Explo a o y Fac o Analysis
ESWA
Expe Sys ems wi h Applica ions
FCUP
Facul y o Sciences o he Uni e si y o Po o
FEUP
Facul y o Enginee ing o he Uni e si y o Po o
FFM
Fi e Fac o Model
FIPA
Founda ion o In elligen Physical Agen s
FIPA-ACL
FIPA Agen Communica ion Language
GDPR
Gene al Da a P o ec ion Regula ion
GECAD
Resea ch G oup on In elligen Enginee ing and Compu ing o Ad anced Inno a ion
and De elopmen
GPS
Global Posi ioning Se ice
GRS
G oup Recommende Sys em
HTTP
Hype ex T ans e P o ocol
IF
Impac Fac o
IPIP
In e na ional Pe sonali y I em Pool
IPP
Poly echnic Ins i u e o Po o
ISCAP
Po o Accoun ing and Business School
ISEP
Supe io Ins i u e o Enginee ing o Po o
JADE
Ja a Agen DE elopmen amewo k
LAN
Local A ea Ne wo k
LASI
In elligen Sys ems Associa e Labo a o y
LBS
Loca ion-Based Se ices
LLM
La ge Language Models
MAMS
Mul i-Agen Mic ose ice
MAS
Mul i-Agen Sys em
MAE
Mean Absolu e E o
MIT
Massachuse s Ins i u e o Technology
ML
Machine Lea ning
Abb e ia ions and Ac onyms
xxi
MS
Mic ose ice
MS-DOS
Mic oso Disk Ope a ing Sys em, o iginally used in IBM pe sonal compu e s
NEO-FFI
Neu o icism-Ex a e sion-Openness Fi e Fac o In en o y
NEO PI-R
Neu o icism-Ex a e sion-Openness Pe sonali y In en o y Re ised
NLP
Na u al Language P ocessing
OCEAN
Openness, Conscien iousness, Ex a e sion, Ag eeableness, Neu o icism
PEGA
Pe sonali y-Guided P e e ence Agg ega o
POI
Poin o In e es
Q&A
Ques ions and Answe s
REST
REp esen a ional S a e T ans e
RMSE
Roo Mean Squa ed E o
RS
Recommende Sys em
SEM
S uc u al Equa ion Modeling
SPMS
Sha ed Se ices o he Minis y o Heal h (Se iços Pa ilhados do Minis é io da
Saúde)
TCCM
Theo y, Con ex , Cha ac e is ics, Me hodology
TIPI
Ten I em Pe sonali y In en o y
UMinho
Uni e si y o Minho
Pa I
EXEGESIS
2
“As long as he cen u ies con inue o un old, he numbe o books will g ow con inually, and one can
p edic ha a ime will come when i will be almos as di icul o lea n any hing om books as om he
di ec s udy o he whole uni e se. I will be almos as con enien o sea ch o some bi o u h con-
cealed in na u e as i will be o ind i hidden away in an immense mul i ude o bound olumes.”
Denis Dide o , “Encyclopédie” (1755)
Chap e 1 – In oduc ion
9
The second s ep, de ini ion o he objec i es o he solu ion, consis ed in he a ionaliza ion o he
objec i es om he ini ially iden i ied p oblems o he GRS p o o ype, being he s a e o he a
ound he base esou ce o ha easoning.
The a i ac s (s ep 3) necessa y o sol e he iden i ied p oblems and ul ill he p oposed objec i es
consis ed in he:
•
concep ualiza ion, design, and de elopmen o se e al models, namely: (1) a model o ca -
ego ize he possible ou is a ac ions, (2) a model o ela e pe sonali y o he p e e ence
o hose ou is a ac ions, (3) ano he model o ela e pe sonali y o a elling mo i a-
ions, and inally, (4) a model o ela e pe sonali y and a el- ela ed p e e ences & con-
ce ns;
•
concep ualiza ion and design o algo i hms o p edic he ou is s’ p e e ences and sol e
he cold-s a p oblem;
•
concep ualiza ion and design o algo i hms o e ec i ely agg ega e ou is s wi h simila
p e e ences in an excu sion g oup and sol e he g oups’ con lic ing p e e ences and he e -
ogenei y;
•
concep ualiza ion, design and de elopmen o he G oup Recommende Sys em p o o ype,
including he implemen a ion o he e e ed models and algo i hms, o p o ide POI ec-
ommenda ions jus by knowing he ou is s’ pe sonali y, and maximize he g oup’s mem-
be s ou is expe ience;
•
de elopmen o algo i hms o imp o e he ini ial ecommenda ions;
•
design o sho -du a ion se ious games as p oo o concep o implici ly de e mine he us-
e s’ pe sonali y.
The p oduced a i ac s a e e lec ed in he pape s ha compose his hesis. Besides being g ound-
ed by he li e a u e e iew, he a i ac s design p ima ily s emmed om da a collec ed om a ques-
ionnai e adminis e ed in a la ge scale s udy (see Pape 2 in Chap e 4, and Pape 3 in Chap e 5).
Demons a ions o he de eloped models and GRS p o o ype o he academic communi y we e
unde aken in di e en i e a ions o he esea ch wo k, showing how he a i ac s sol ed he iden i-
ied issues, ep esen ed by Pape s 4 and 6 (Chap e 6 and Chap e 8, espec i ely).
Func ional and in eg a ion es s along he de elopmen p ocess de e mined he need o i e a e
back o s ep 3 o y o imp o e he models, algo i hms, and he GRS p o o ype. The p o o ype was

Chap e 1 – In oduc ion
10
also e alua ed o pe o mance, scalabili y, sensi i i y, accu acy, and use sa is ac ion a e i s de-
elopmen , co esponding o DSRM s ep 5, and is p esen ed in Pape 5 (Chap e 7). This included
expe imen a ion and case s udies wi h eal use s. In ano he i e a ion, he se ious games de el-
oped we e also es ed and e alua ed in expe imen s wi h eal use s, which is add essed in Pape 7
(Chap e 9).
The communica ion s ep accompanied he whole p ocess o he doc o al wo k, which in ol ed
communica ing he impo ance o he iden i ied p oblems and p oposed solu ions, he u ili y, no el-
y and e ec i eness o he designed and de eloped a i ac s o o he schola s, h ough he w i ing
o he men ioned scien i ic pape s, published and publishable, in he a eas o Use Modeling, Hu-
man-Compu e In e ac ion and Recommende Sys ems.
The wo k p esen ed in his hesis was conduc ed a he Resea ch G oup on In elligen Enginee ing
and Compu ing o Ad anced Inno a ion and De elopmen (GECAD), esea ch cen e o he Supe-
io Ins i u e o Enginee ing o he Poly echnic o Po o (ISEP, IPP), Po ugal, in associa ion wi h he
ALGORITMI Resea ch Cen e, a esea ch uni o he School o Enginee ing o he Uni e si y o Mi-
nho (UMinho), B aga, Po ugal. Due o he CoViD-19 pandemics, pa o he wo k was conduc ed
emo ely. This doc o al wo k was suppo ed by he FCT esea ch g an 2020.06129.BD and by he
Eu opean Social Fund. Be o e being awa ded he FCT g an , I was a esea ch ellow o he
G ouplanne p ojec (unde g an POCI-01-0145-FEDER-29178). As some o he doc o al wo k was
used o he G ouplanne p ojec , he GRS p o o ype adop ed he “G ouplanne ” name.
1.4
T
HESIS
S
TRUCTURE
This hesis is based on a compila ion o published and publishable pape s, also known as “compi-
la ion hesis” o “a icle-based hesis” (Gus a ii, 2012; Mo ison, 2017; Rowe, 2015). I ollows he
“sandwich o ma ” (Menk Dos San os, 2019) and comp ises h ee pa s: Pa I, he exegesis,
which is composed o wo chap e s, his in oduc ion and a con ex ualiza ion on he mos ele an
add essed opics; Pa II, he compila ion o he pape s ha includes 6 published pape s, 2 in
jou nals and 4 in con e ences, and ano he pape , cu en ly in e iew, submi ed o a jou nal, co -
esponding o se en chap e s in o al; and inally, Pa III, which co esponds o he main conclu-
sion, limi a ions ound and u u e di ec ions.
Chap e 1 – In oduc ion
11
In Pa II, he se en w i en pape s a e p esen ed as is, cons i u ing each one a chap e , s a ing in
Chap e 3. Fo eadabili y, he pape s’ lis o e e ences we e compiled a he end o he hesis in a
global lis o e e ences (Gus a ii, 2012; Menk Dos San os, 2019), as well as he espec i e ap-
pendices. The se en pape s included we e p oduced du ing he se e al s ages o he esea ch
p ojec and cons i u e he de ailed main s o yline o he doc o al wo k. The concep s, ideas, and
p ocess behind he pape s a e sho ly desc ibed nex .
Pape 1. Modeling a Mobile G oup Recommende Sys em o Tou ism wi h In elligen
Agen s and Gami ica ion
This was he i s pape w i en o he hesis p ojec and con ibu es o objec i es O1 and O2. I
ep esen s he i s ideas hough up o he GRS p o o ype which I was h illed o s a idealizing.
This i s concep was impo an o b ains o ming and de e mining he conduc ing wi e o he
p oposed esea ch. I in ol ed he i s opics add essed in he wo k, such as o be e unde s and
RS, wha al eady exis ed in li e a u e o add ess ecommenda ions o g oups, and especially o
g oups o ou is s.
A ho ough li e a u e e iew on RS and GRS was conduc ed, wha aspec s should be conside ed,
and wha he main limi a ions we e. A e y impo an and ac ual p oblem ela ed o RS was
acknowledged o he i s ime, he cold-s a p oblem. This highligh ed he pa h by un eiling he
impo ance o pe sonali y in p edic ing use p e e ences, leading o he conclusion ha mo e in-
o ma ion on pe sonali y was needed.
Du ing he esea ch, i was ound ha game componen s we e being used o imp o e he use s
in e ac ion in non-game applica ions, lea ning he e m gami ica ion, which I immedia ely ound a
e y in e es ing opic due o my in ol emen in he ideogames wo ld. This in oduced he idea o
how i could be used o imp o e a RS, as i was shown o be a success among s uden s, wo ke s,
and o he s (de CA Zieseme , Mülle , & Sil ei a, 2014; Hama i e al., 2014). Howe e , some ea-
u es idealized in he pape ended up being p ojec ions o u u e p ojec s ou side he scope o he
cu en wo k, such as he use ’s a a a o i ual pe , and he use o Augmen ed Reali y (AR). The
gami ica ion o he p o o ype was ini ially planned bu was la e emo ed om he scope o he
doc o al wo k due o he complexi y in ol ed.
Chap e 1 – In oduc ion
12
Wi h he in o ma ion ga he ed, he GRS s a ed o be idealized, and he i s concep ual model was
designed. This s udy opened he pa h needed o de e mine wha ac o s would be mo e impo an
o imp o e g oup ecommenda ions, and wha psychological aspec s would be ele an o pe son-
alize ecommenda ions.
Pape 2. Modeling Tou is s' Pe sonali y in Recommende Sys ems: How Does Pe -
sonali y In luence P e e ences o Tou is A ac ions?
The wo k in ol ed in he w i ing o his pape was e y challenging and inspi ing, as i also in ol ed
esea ch in he Psychology a ea. I was a e y impo an esea ch and de e minan o he doc o al
wo k. I in ol ed ex ensi e esea ch on wha pe sonali y was, how i could be cha ac e ized, and
he main models used in li e a u e, and con ibu es o objec i e O3, and esea ch ques ions
RQ1.1, RQ1.2, and RQ2.
A ho ough li e a u e e iew on he in luence o pe sonali y in he use s’ p e e ences in a a ie y o
domains was pe o med, especially in he ou ism indus y, and he use o pe sonali y in RS was
explo ed. I also lea ned he e m Psychology o Tou ism and go ascina ed by he a ea. This ook
o a challenge, ela ed o esea ch ques ions RQ1.1 and RQ1.2. How could he ela ionship be-
ween he aw i e pe sonali y dimensions and he ou is s’ p e e ences be de e mined, ins ead o
using ou is ypologies like he mos exis ing wo ks?
Del ing he mos impo an e ms used o classi ying ou is a ac ions, he mos ele an e ms
we e chosen in b ains o ming mee ings wi h wo co-au ho s, one o hem a psychologis and
G ouplanne ’s esea ch ellow, Ped o Sa ai a, and an on ology s a ed o be idealized. The e ms
chosen we e d awn in a mind map, which was named he “Fish On ology” due o i s shape and
can be consul ed in Appendix A. Then, o ind he ela ionship be ween ou is p e e ences and he
ou is s’ pe sonali y, a su ey o ga he g ound- u h da a was conduc ed, which was also challeng-
ing. The su ey in ol ed ha d wo k and pe sis ence, as he ques ionnai e comp ised 196 ques-
ions, and no e e yone was willing o answe o ee
3
.
To de ine he “pe sonali y s ou is p e e ences” model, many weeks o model adjus men s and
calcula ions we e needed. As a esul , we cons uc ed a model ha p o ed wha I expec ed, ha
3
The ques ionnai e can be consul ed a h ps://www.gecad.isep.ipp.p /g ouplanne /dissemina ion.h ml, unde he “UMUAI APPENDIX” pane.
Chap e 1 – In oduc ion
13
he aw pe sonali y dimensions could be used o p edic ou is p e e ences. Howe e , he model’s
i was only sa is ac o y, meaning he sample size needed o be la ge o inc ease he model’s i ,
which was al eady being idealized as u u e wo k (see Pape 3).
Ha ing his pape accep ed in one o he op con e ences on he use modelling a ea a he ime
was e y g a i ying as i ecognized he impo ance o he wo k done.
Pape 3. G oup ecommende sys ems o ou ism: how does pe sonali y p edic
p e e ences o a ac ions, a el mo i a ions, p e e ences and conce ns?
This pape con inues he wo k p esen ed in Pape 2, also ocusing in answe ing o esea ch ques-
ions RQ1, RQ1.1 and RQ1.2, and objec i es O3 and O4, by imp o ing he p e iously ob ained
“Pe sonali y s Tou is A ac ions P e e ence” model wi h a la ge and mo e he e ogenous sam-
ple, and ex ending he p edic ion o a elling mo i a ions, and a el- ela ed p e e ences & con-
ce ns, as we also in ended o conside mo e pe sonal conce ns and in e es s, including wea he
p e e ence, ea s/phobias and physical limi a ions, o maximize he ou is s’ expe ience.
The esea ch he e p esen ed esul s om a long and exhaus i e s udy ( om 2020 o 2022)
4
o-
cused in disco e ing how pe sonali y ela es o he choice o (a wide ange o ) ou is a ac ions,
a elling mo i a ions, and a el- ela ed p e e ences & conce ns, as mos o he wo ks ound in
li e a u e we e based on ou is oles/ ypologies and/o do no use he Big Fi e
5
pe sonali y di-
mensions
o p edic objec i e ela ionships be ween pe sonali y and he e e ed a el aspec s.
The ques ionnai e om he p e ious s udy was used, and, wi h many pe sis ence, he sample was
duplica ed, esul ing in a o al o
𝑛
=1035 iable esponses.
In his wo k, objec i e ela ionships be ween all he Big Fi e pe sonali y dimensions, and some
espec i e ai s, and he p e e ence o a wide ange o ou ism ca ego ies, a elling mo i es,
a el- ela ed p e e ences & conce ns, and wea he p e e ences we e ound, con ibu ing wi h
knowledge and hope ully a solid base o esea che s o RS o ou ism o au oma ically model
4
Whe e I had he pleasu e o exchanging messages wi h Paul Cos a, ega ding some doub s abou he Re ised NEO pe sonali y in en o y (NEO PI-R)
(Cos a & MacC ae, 1992). Thank you o you sympa hy and sugges ions on wo k o ollow.
5
Model mos widely ecognized o ep esen he i e pe sonali y dimensions (Dhelim, Aung, Bou as, Ning, & Camb ia, 2021; Digman, 1990; Ma z, Chan, &
Kosinski, 2016).
Chap e 1 – In oduc ion
14
ou is s based on hei pe sonali y. The esul s we e used o p opose ecommenda ion algo i hms
o be used in he GRS p o o ype.
These indings allow o sol e he cold-s a p oblem and c ea e g oups wi h simila p e e ences,
and he e o e less con lic ing p e e ences, so mo e accu a e and pe sonalized ecommenda ions
can be p o ided, being, o he bes o ou knowledge, he i s s o p o ide an exhaus i e p edic ion
o a el aspec s p e e ence based on he ou is s’ aw pe sonali y, also con ibu ing o esea ch
ques ions RQ3 and RQ4.
This pape was published in one o he op jou nals in he use modeling a ea a he ime and is
p esen ed in Pa II.
Pape 4. G ouplanne : A G oup Recommende Sys em o Tou ism wi h Mul i-Agen
Mic ose ices
Meanwhile, he GRS p o o ype concep ini ially p oposed was imp o ed, and s a ed o be imple-
men ed a he beginning o 2021. This in ol ed ho ough equi emen s elici a ion and p o o yping,
he a chi ec u e and domain model design, and he design and implemen a ion o pe sonali y-
based ecommenda ion algo i hms, conside ing he indings o Pape s 2 and 3.
In Feb ua y o 2022, he p o o ype s a ed o be al e ed, whe e he main change was o ans o m
he MAS a chi ec u e in o a Mul i-Agen Mic oSe ice (MAMS), joining he bes o a Mul i-Agen
Sys em and Mic ose ices, whe e he agen s would be di ec ly accessed h ough Rep esen a ional
S a e T ans e (REST) endpoin s. So, he a chi ec u e model p oposed in Pape 1 was imp o ed
and is b ie ly p esen ed in his demo pape
6
and mo e comple ely in Pape 5.
This pape con ibu es o objec i es O1, O2, and O4, and esea ch ques ions RQ2 and RQ3.
This demons a ion was e y impo an o he esea ch wo k and o me pe sonally, being awa d-
ed he bes demo pape p ize o he IBM Demons a ion Awa d a PAAMS’22, which was a ecogni-
ion o he wo k done so a .
6
The demons a ion ideo can be wa ched a h ps://www.you ube.com/wa ch? =sS3Zc9k 0NA

Chap e 1 – In oduc ion
15
Pape 5. A e He e ogeini y and Con lic ing P e e ences No Longe a P oblem? Pe -
sonali y-Based Dynamic Clus e ing o G oup Recommende Sys ems
The algo i hms o he ecommenda ions and o ma ion o subg oups p oposed in Pape 3 we e
ho oughly analyzed and signi ican ly imp o ed, being implemen ed acco dingly in he GRS p o o-
ype. The poin s o in e es a ings we e no longe needed o p o ide he i s ecommenda ions, as
he ecommenda ions we e solely based on he ou is s’ pe sonali y, by using he pe sonali y mod-
els p oposed in Pape 3 o p edic he a el p e e ences, sol ing he cold-s a p oblem. To u he
pe sonalize he ecommenda ions and show conce n o he ou is s’ speci ic needs, ea s/phobias
and disabili ies, i any, we e also conside ed and implemen ed in he ecommenda ion algo i hms.
A new pe sonali y-based dynamic clus e ing algo i hm,
𝑑
-means, was also concep ualized and
implemen ed in he MAMS, by adap ing he
𝑘
-means algo i hm, so he ou is s could be agg ega -
ed acco ding o hei pe sonali y using a high simila i y h eshold (≥ 0.80). This adap a ion was
easoned o elimina e he need o know he numbe o clus e s a p io i, as is manda o y by he
exis ing clus e ing algo i hms, o elimina e noise and ou lie s, which a e a common p oblem in
clus e ing algo i hms, and o sol e g oup’s con lic ing p e e ences and he e ogeini y. This is a no -
el y in he RS a ea, assigning each ou is o a clus e in eal- ime wi hou needing o know he
numbe o clus e s be o e.
The idea o using associa ion ules, namely he Ap io i algo i hm, o u he p edic he clus e ed
ou is s’ p e e ences came up. This idea was designed o help imp o e and e ine he ini ial ec-
ommenda ions by he Recommenda ion Engine, using he ou is s’ p o ile and a el his o y.
To also imp o e he ecommenda ions, he wea he condi ions a he excu sions’ des ina ion and
da es we e also conside ed, using a wea he API, p o ided by Sis ade a he ime, as we we e
join ly wo king in a new p ojec , he sma T a el p ojec
7
.
The wo k he e p esen ed con ibu ed o esea ch ques ions RQ2, RQ3, and RQ4, and objec i es
O1 o O5. To ha e his wo k accep ed by Expe Sys ems wi h Applica ions is e y ewa ding and
mo i a ing, and ep esen s an impo an miles one in my doc o a e.
7
Mo e in o ma ion a h ps://sma a el.sis ade.com/
Chap e 1 – In oduc ion
16
Pape 6. Imp o ing G oup Recommenda ions using Pe sonali y, Dynamic Clus e ing
and Mul i-Agen Mic ose ices
This pape ep esen s a demons a ion o he wo k published in Pape 5, p esen ed a he 17 h
ACM Con e ence on Recommende Sys ems (RecSys’23), and con ibu ed o objec i es O1 o O5,
and esea ch ques ions RQ2, RQ3, and RQ4.
RecSys is he p emie in e na ional o um o he p esen a ion o wo ks ela ed o RS. “I has be-
come he mos impo an annual con e ence o he p esen a ion and discussion o ecommende
sys ems esea ch” (RecSys, 2023), and is sponso ed by g ea companies like Ne lix Resea ch,
Huawei, Amazon Science, Google, Me a, among o he s.
A pos e ep esen ing he wo k (see Appendix C) as well as a demons a ion ideo
8
we e c ea ed
o he con e ence
9
.
This con e ence had been in my mind since he beginning o my doc o a e and was ano he mile-
s one I eally wan ed o accomplish due o he con e ence’s epu a ion in he RS a ea. To ha e his
demo accep ed was he e o e a e y impo an conques o me.
Pape 7. "You Wan o Play a Game?" De ec ing Two Pe sonali y T ai s wi h Sho -
Du a ion Mobile Games.
This is he las pape ha composes his hesis and is cu en ly unde e iew in an in e na ional
jou nal. I p esen s he wo k ela ed o he concep o using se ious games o cap u e he ou is s’
pe sonali y, as I no me ely wan ed o y o cap u e he ou is s’ pe sonali y implici ly, bu equally
he mos g anula Big Fi e pe sonali y ai s. No only o c ea e a mo e imme si e, mo i a ing and
un way o cap u ing he ou is s’ pe sonali y, bu also o sol e he p oblems ela ed o sel -
epo ing ques ionnai es, such as he social desi abili y bias and ake esponses.
As p oo o concep , I chose o design wo games, each one in ended o speci ically acqui e one o
he 30 Big Fi e ai s (Cos a & MacC ae, 1992). In a b ains o ming session wi h wo o he co-
au ho s, João T indade and Gonçalo Mon ei o, also my men ees and sma T a el’s ellow s u-
8
The ideo can be consul ed a h ps://you u.be/up0A3B_e6zU
9
The GRS p o o ype use manual can be consul ed a h ps://www.gecad.isep.ipp.p /g ouplanne /asse s/ iles/g ouplanne _use _manual.pd
Chap e 1 – In oduc ion
17
den s
10
, who we e also used o play di e en ypes o ideo and mobile games, we decided o
cau iousness and achie emen -s i ing, as hey we e ela ed o he Conscien iousness dimension,
which we ind e y in e es ing, and because hey seemed o be easie o measu e in a i s con-
cep .
Howe e , he design o he games was de ying, whe e he cau iousness ai e ealed o be he
mos complex o acqui e, as no many ela ed wo ks could be ound in li e a u e, in ol ing mo e
imagina ion. Se e al si ua ions ha we hough could cause he playe s o beha e in a mo e ca e-
ul way we e designed, inspi ed by se e al games played in he pas , like P ince o Pe sia and
Tomb Raide , among o he s. The idea o he achie emen -s i ing game was also inspi ed in clas-
sic pla o me s, like P ehis o ik, Sonic, and Supe Ma io B os, and mo e ecen games like G an
Tu ismo o he ophies’ mechanics.
The game design ideas we e deba ed in o he b ains o ming sessions, including Ped o Sa ai a,
who also enjoys playing ideo and mobile games, con ibu ing knowledge om he beha io al a ea
o Psychology. To de e mine i he minigames we e measu ing he in ended pe sonali y ai s, ex-
pe imen s wi h eal use s we e conduc ed (
𝑛
=100).
I can say ha s udying and designing he games was e y g a i ying and un o me, as I ha e al-
ways enjoyed playing ideogames since my childhood.
In his pape , we show pe sonali y ai s can be measu ed wi h sho -du a ion mobile se ious
games, and ha being able o measu e pe sonali y ai s o he han he p oposed ones is a plus.
This also shows ha i migh no be possible o measu e some ai s independen ly, as hey a e
ela ed o each o he . This means ha wi h simple minigames, wi hou many in e ac ions, we can
measu e someone’s de ailed pe sonali y wi hou being subjec ed o he social desi abili y bias and
alse esponses, in less han i e minu es. O cou se, he e is always he possibili y o illing in he
pe sonali y ques ionnai es, bu we al eady know ha many esponses can be socially biased o
alse, which would hinde he cha ac e iza ion o he eal ue sel . The p esence o a hi d pe son
o conduc he ques ionnai e would be ime-consuming and logis ically impossible o modeling all
use s in an RS. This pape con ibu es o objec i e O3 and esea ch ques ion RQ1.
10
sma T a el P ojec suppo ed unde he Eu opean Regional De elopmen Fund POCI-01-0247-FEDER-179946.
18
2
M A I N C O N C E P T S
To gi e some ini ial backg ound o he eade , being ca e ul no o epea he con en add essed in
he pape s w i en o he doc o al wo k, his Chap e summa izes he main concep s behind he
wo k done, such as wha is a Recommende Sys em, wha a e he main ypes, and some cu en
key limi a ions. G oup Recommende Sys ems a e also p esen ed, as well as he concep s o Mul i-
Agen Mic ose ices, Pe sonali y, Gami ica ion, and Se ious Games. De ailed in o ma ion, con ex-
ualiza ion and s a e o he a is gi en in he pape s ha compose Pa II.
2.1
R
ECOMMENDER
S
YSTEMS
A Recommende Sys em (RS) is a ype o in o ma ion il e ing ool designed o sugges i ems ha
a e mos ele an o a speci ic use (Ricci, Rokach, & Shapi a, 2015), and a e cu en ly used in a
panoply o domains, like online s o es, music, mo ies, es au an s, ou ism, online da ing, e c. A
RS ypically specializes in a pa icula ype o i em, whe e he e m "i em" e e s o wha he sys-
em sugges s o use s, such as news o ead, a place o isi , o music o lis en (see Figu e 2.1
Le ). As a esul , i s design, use in e ace, and he main ecommenda ion echnique i employs
a e all ailo ed o o e use ul and e ec i e sugges ions o ha speci ic ca ego y o i ems. These
sys ems a e especially help ul when a pe son needs o selec an i em om a la ge and po en ially
o e whelming numbe o op ions p o ided by a se ice, a p oblem known as in o ma ion o e load
(Le y, 2008), and al eady o eseen by Denis Dide o in 1755 (Dide o & d'Alembe , 1776).
This p oblem was compu a ionally app oached o he i s ime by Elaine Rich (Rich, 1979), who
de eloped he i s RS, G undy, a i ual lib a ian ha sugges ed books o indi idual use s acco d-
ing o hei s e eo ype. Al eady a ha ime, she s a ed ha compu e s could dis inc i ely ea us-
e s acco ding o hei pe sonali y and objec i es, and, he e o e, p oposed o model use s acco d-
Chap e 2 – Main Concep s
25
Con en -Based Fil e ing:
This app oach sugges s i ems o use s based on he cha ac e -
is ics o he i ems and he use 's pas in e ac ions o p e e ences. I elies on he a ibu es
o i ems (like gen e, desc ip ion, keywo ds) and compa es hem o wha he use has p e i-
ously liked o in e ac ed wi h. Fo example, in ou ism, con en -based ecommende s migh
sugges a ac ions like ones he use has isi ed be o e.
Collabo a i e Fil e ing:
Collabo a i e il e ing was coined o he i s ime by D. Goldbe g
e al. (1992). This app oach ecommends i ems based on he p e e ences and beha io s o
o he use s. I ope a es unde he assump ion ha i use s ag eed on i ems in he pas , hey
would likely ag ee again in he u u e. Fo example, i wo use s ha e a ed se e al places
simila ly, he sys em migh ecommend a place one o hem liked o he o he . Collabo a i e
il e ing is di ided in o wo main ypes: use -based il e ing, which ecommends i ems o a
use by inding simila use s (neighbo s) who ha e simila as es o p e e ences; and i em-
based il e ing, which ecommends i ems by inding simila i ems based on he p e e ences
o all use s, i.e., inds i ems ha a e like hose he use has p e iously liked o a ed highly
and ecommends hose. Amazon was one o he pionee s in implemen ing ecommenda ion
echnology on a la ge scale (Linden, Smi h, & Yo k, 2003).
Collabo a i e ecommenda ion algo i hms can be u he ca ego ized in o wo b oad ypes:
memo y-based (also known as heu is ic-based), which use he en i e use -i em a ings da a-
base o make p edic ions; and model-based, which in ol e c ea ing/lea ning a model om
he use -i em in e ac ions (e.g., using machine lea ning echniques like ma ix ac o iza ion,
clus e models, Bayesian ne wo ks, and deep lea ning echniques) o make p edic ions
(B eese, Hecke man, & Kadie, 2013).
Hyb id App oaches:
These sys ems combine mul iple ecommenda ion echniques, such
as con en -based and collabo a i e il e ing, o le e age he s eng hs o each o he and p o-
ide mo e accu a e and obus ecommenda ions, and o he machine lea ning echniques,
such as deep neu al ne wo ks like in YouTube (Co ing on, Adams, & Sa gin, 2016), deep
lea ning, na u al language p ocessing, and compu e ision like in Ne lix (Bansal & Sha ma,
2024). Hyb id ecommende sys ems can be enhanced wi h knowledge-based echniques,
such as case-based easoning, o imp o e ecommenda ion accu acy and o o e come some

Chap e 2 – Main Concep s
26
o he limi a ions, such as issues wi h new use s and new i ems, aced by adi ional ec-
ommende sys ems, as unlike collabo a i e o con en -based me hods, knowledge-based
echniques do no ely on his o ical use da a. Ins ead, hey use explici knowledge abou us-
e needs and p e e ences and i em ea u es o gene a e ecommenda ions, which is use ul
o domains whe e use p e e ences a e highly speci ic o change in equen ly, like in he
ou ism, eal es a e, o au omobiles domains (Agga wal, 2016). As poin ed by Smy h (2007)
and e e ed by Agga wal (2016), knowledge-based RS a e equen ly seen as closely ela ed
o con en -based RS, leading o ques ions abou whe he a clea dis inc ion exis s be ween
he wo ypes o me hods.
Fu he de ails on he di e en ypes o RS can be consul ed, o example, in he RS handbooks o
Adoma icius and Tuzhilin (2005), Ricci e al. (2015), and Agga wal (2016).
I is clea ha RS ha e signi ican ly e ol ed since hei i s appea ance, and ha many issues a e
s ill open, such as da a spa si y and he cold-s a p oblem, whe e he lack o su icien da a on
new use s o new i ems can make i di icul o he sys em o p o ide accu a e ecommenda ions.
Each RS ype has i s own s eng hs and weaknesses, and he app oach o use o en depends on
he speci ic use case and he na u e o he a ailable da a.
Conside ing he exis ing ypes o RS and he issues iden i ied, a hyb id ecommende sys em p o-
o ype was concep ualized and de eloped o he doc o al wo k. I uses use -based collabo a i e
il e ing wi h demog aphic il e ing, con ex -awa eness, knowledge-based o some ex en , and mod-
el-based algo i hms such as clus e ing and associa ion ules. The de elopmen p ocess is de-
sc ibed h oughou he pape s in Pa II, along wi h p ope jus i ica ions.
2.2
P
ERSONALITY
-B
ASED
R
ECOMMENDER
S
YSTEMS
Th oughou he las wo decades, pe sonaliza ion became he co e ea u e o he e ec i eness o
RS in a wide ange o domains (L. Chen, Wu, & He, 2016; Rashid e al., 2002; Tkalcic & Chen,
2015; Tondello, O ji, & Nacke, 2017), wi h se e al ying o p edic he use s’ p e e ences based
on hei pe sonali y (Dhelim e al., 2023; Eldeswky, Elazab, Bolock, & Abdennadhe , 2023; Hu &
Pu, 2009; Pe ik, De Ruy e , Ma kopoulos, & Eggen, 2004; Roshchina, 2012; Tkalcic & Chen,
2022; Tkalcic, Kuna e , Tasic, & Koši , 2009; W. Wu, 2017). Nunes, Ce i, and Blanc (2008) p o-
Chap e 2 – Main Concep s
27
posed he use o pe sonali y ai s o ecommend a p esiden candida e o o e s in he elec ions
in F ance. In an expe imen , he pa icipan s had o e alua e wo candida es and he ideal p esi-
den using he IPIP-NEO-300 pe sonali y ques ionnai e. Compa ing he esul s o he pa icipan s
ac ual o es, when using he 30 pe sonali y ai s he ecommenda ions we e 100% accu a e.
When using only he 5 pe sonali y dimensions, he ecommenda ions we e 80% accu a e.
W. Wu, Chen, and Zhao (2018) used he BFI o assess he use s’ di e si y p e e ence o di e en
aspec s o li e’s opics in a popula Chinese social ne wo k, such as music, mo ies, spo s, e c.
Pe sonali y simila i y was hen used in collabo a i e il e ing and g eedy e- anking echniques o
make ecommenda ions o opics.
Dhelim, Ning, Aung, Huang, and Ma (2020) de eloped Me a-In e es , a pe sonali y-based p oduc
ecommende sys em. They combined da a-mining echniques, such as ex -mining in social-
ne wo ks, wi h pe sonali y and p e e ences simila i y o ma ch he i ems o ecommend in a news
social ne wo k. They used he sho e sion o he Big Fi e In en o y (BFI), he Ten I em Pe sonali y
In en o y (TIPI) (Gosling, Ren ow, & Swann J , 2003), o de e mine he use s’ pe sonali y.
PEGA is a Pe sonali y-Guided P e e ence Agg ega o model used o making ecommenda ions o
epheme al g oups, p oposed by Ye e al. (2023). They use he use s’ indi idual Big Fi e pe sonali-
y o calcula e he weigh ed g oup p e e ences and make he ecommenda ions acco ding o hose
weigh s.
Many o he pe sonalized RS ha conside he use s’ pe sonali y ha e been de eloped (Dhelim e
al., 2021; Wa is, Fakha , Gulsoy, Yalcin, & Bilge, 2024). Howe e , o he bes o ou knowledge,
he exis ing RS ob ain he use s’ pe sonali y om explici pe sonali y ques ionnai es, o implici ly
by lea ning om hei in e ac ions wi h he con inuous use o he sys em, like da a mining social
cha s o pos s. Tha is, hey ha e ime-consuming con igu a ions and/o excessi e in usi eness,
o hey need o da a mine a g ea amoun o in o ma ion in o de o lea n he use ’s beha io and
in e es s. Also, he measu ed pe sonali y is based on he b oades i e dimensions ins ead o he
mo e g anula ai s, and some use less accu a e ques ionnai es, such as he TIPI. To o e come
hese limi a ions, as a p oo o concep , we de eloped sho -du a ion se ious games o implici ly
de e mine g anula pe sonali y ai s, which a e desc ibed in he wo k composing Pape 7 (Chap e
9).
Chap e 2 – Main Concep s
28
2.3
G
ROUP
R
ECOMMENDER
S
YSTEMS
G oup ecommende sys ems a e ools designed o sugges i ems o a g oup o use s a he han
jus a single indi idual. These sys ems conside he p e e ences o all membe s in he g oup o
gene a e a ecommenda ion aiming o be e sa is y he g oup as a whole. The goal is o ind a
balance be ween indi idual p e e ences o p o ide sugges ions ha a e accep able, o e en enjoy-
able, o he en i e g oup.
GRS eme ged as a na u al ex ension o pe sonalized RS, which ha e been widely used since he
mid-1990s wi h he ise o pla o ms like Amazon and Ne lix, whe e o p o ide pe sonalized ec-
ommenda ions was essen ial o helping use s na iga e as amoun s o con en (Jannach e al.,
2021). Howe e , i soon became e iden ha many eal-wo ld si ua ions in ol e g oups o people
making join decisions, such as iends deciding on a mo ie o amilies choosing a aca ion des i-
na ion (Jameson, Baldes, & Kleinbaue , 2003).
The ea lies esea ch on GRS began in he ea ly 2000s. One o he pionee ing s udies was Pol-
yLens (O’conno , Cosley, Kons an, & Riedl, 2001), a GRS which adap ed pe sonalized ecommen-
da ion echniques o ca e o g oup p e e ences. This wo k eme ged as pa o a b oade e o o
ex end collabo a i e il e ing echniques, which we e ini ially de eloped o indi idual use s, o
g oup con ex s. Many ounda ional concep s om pe sonalized RS, like collabo a i e il e ing and
p e e ence agg ega ion, we e adap ed and expanded o c ea e algo i hms ha could accommoda e
mul iple use s' p e e ences (Jameson & Smy h, 2007; Mas ho , 2003). Since hen, he ield has
g own signi ican ly, wi h applica ions spanning en e ainmen , a el, e-comme ce, and e en col-
labo a i e wo k en i onmen s (Delic & Mas ho , 2018; Mas ho , 2015).
While GRS ha e made signi ican ad ancemen s, hey s ill ace se e al challenges (Ál a ez
Má quez & Ziegle , 2016; Delic & Mas ho , 2018; Delic, Neidha d , Nguyen, & Ricci, 2018; Delić
e al., 2020; Ricci, 2022). The mos undamen al issue is how o combine he p e e ences o indi-
idual g oup membe s. This can be done using me hods like a e aging a ings o selec ing he
mos popula choice wi hin he g oup. Howe e , hese me hods o en ail o cap u e he nuances o
di e se p e e ences, leading o subop imal esul s o ce ain g oup membe s. The na u e o g oup
in e ac ions is complex, including nego ia ion, comp omise, and he in luence o dominan pe son-
ali ies, whe e g oup membe s in luence each o he ’s p e e ences and choices, and unde s anding
Chap e 2 – Main Concep s
29
hese dynamics is c ucial o de eloping e ec i e GRS (Delic e al., 2018). This gi es ise o ano h-
e challenge, ensu ing ai ness and balancing he sa is ac ion o all g oup membe s. Simply selec -
ing a "majo i y ule" app oach can aliena e mino i y p e e ences wi hin he g oup. Designing algo-
i hms ha gi e oice o all membe s wi hou disp opo iona ely a o ing any one indi idual e-
mains a key issue. The use s’ sa is ac ion, no jus wi h he inal ecommenda ion bu wi h he
choice-making p ocess i sel , is e y impo an , which is no easy o p edic . Fo example, he con-
side a ion o cogni i e and emo ional aspec s has been shown o imp o e he accu acy in he sa is-
ac ion e alua ion and p edic ion (João Ca nei o, Sa ai a, e al., 2019).
I is also common, and e en mo e in la ge (occasional) g oups, he exis ence o he e ogeini y and
con lic ing p e e ences among he g oup membe s. Fo ins ance, one pe son migh wan o isi a
monumen , while ano he p e e s o do wild na u e ac i i ies. Resol ing hese con lic s in a way
ha main ains ha mony and p o ides an ag eeable ecommenda ion is di icul . Also, as he num-
be o g oup membe s inc eases, he complexi y o calcula ing a ecommenda ion ha balances
e e yone’s p e e ences g ows. This becomes compu a ionally demanding, especially in la ge
g oups.
The composi ion o g oups o en changes o e ime. A sys em ha wo ks well o a g oup o iends
may need o adjus when one membe is eplaced o a new pe son joins in. Handling hese dy-
namic shi s in g oup composi ion while main aining ele ance in ecommenda ions is an ongoing
challenge.
G oup choices o en depend hea ily on con ex . Fo example, he same g oup migh p e e di e -
en es au an ecommenda ions depending on whe he i ’s a casual nigh ou o a special occa-
sion. Designing sys ems ha can accoun o such con ex -speci ic p e e ences is complex and
equi es in eg a ing con ex ual da a.
In eal-wo ld scena ios, g oup p e e ences may e ol e as discussions un old. Fo ins ance, a e
some nego ia ion, g oup membe s migh so en hei ini ial p e e ences o change hei minds en-
i ely. GRS need o adap in eal- ime o hese changes, which p esen s a signi ican echnical chal-
lenge, al hough i can be a con o e sial subjec whe he he g oup’s membe s should be in lu-
enced o no by he o he s (Delic & Mas ho , 2018; Delic, Neidha d , Rook, We hne , & Zanke ,
2017).
Chap e 2 – Main Concep s
30
GRS ep esen an exci ing and necessa y e olu ion in ecommenda ion echnology, bu hey s ill
ace impo an challenges. While we ha e come a long way since hei incep ion in he ea ly
2000s, challenges ela ed o p e e ence agg ega ion, ai ness, and eal- ime adap abili y emain
a eas o ac i e esea ch and de elopmen . As hese sys ems con inue o imp o e, hey hold g ea
po en ial o enhance choice-making o g oups in a wide ange o con ex s.
2.3.1
Examples o GRS in Indus y
Se e al GRS p o o ypes, in di e en domains, can be ound in li e a u e, being INTRIGUE
(A dissono e al., 2003), T a el Decision Fo um (Jameson e al., 2003), CATS (McCa hy, Salamó,
e al., 2006a), and e-Tou ism some o he i s (Sebas ia, Ga cia, Onaindia, & Guzman, 2009) in
he ou ism domain. The GRS p oposed by Sojah ood, Taleai, and Cheng (2023), Abolghasemi,
Khadka, e al. (2022), and CHARM, de eloped by Delic e al. (2024), a e some examples o ecen
ones. These and o he wo ks a e de ailed in he pape s ha compose his hesis.
Se e al indus ies ha e also implemen ed GRS o ca e o he collec i e p e e ences o mul iple
use s. Fo ins ance, Spo i y o e s a g oup session ea u e, Spo i y Jam sessions
11
(p e iously
Spo i y Blend), whe e use s can me ge hei lis ening habi s wi h iends o amily o c ea e collab-
o a i e playlis s o enjoy sha ed sessions in eal- ime.
Facebook E en s
12
inco po a es elemen s o g oup ecommenda ion, pa icula ly when sugges ing
loca ions o ac i i ies o social ga he ings. The pla o m uses da a om g oup membe s, such as
liked e en s, p e ious ac i i y, o in e ac ions wi h simila pos s, o ecommend e en s o places
ha migh in e es he en i e g oup.
Bo h T ipAd iso and Ai bnb o e ecommenda ions o g oups a eling oge he , p o iding sug-
ges ions o accommoda ions, ac i i ies, and dining. Fo example, when mul iple use s inpu hei
p e e ences, he pla o ms can ecommend aca ion en als o ou is a ac ions ha sui he col-
lec i e as es o he g oup.
11
Mo e de ails on h ps://communi y.spo i y.com/ 5/You -Lib a y/Spo i y-Takes-Lis ening-Wi h-F iends- o-Ano he -Le el-wi h-Jam/ d-p/5641808
12
h ps://www. acebook.com/e en s/

Chap e 2 – Main Concep s
31
Google Maps’ G oup Planning
13
ea u e allows use s o sha e lis s o places hey wan o isi (e.g.,
es au an s o ou is ic si es) and collec i ely o e on which place o go. The sys em ecommends
places based on sha ed p e e ences wi hin he g oup.
Booking.com uses collabo a i e il e ing and g oup ecommenda ion echniques o sugges ac-
commoda ions ha sui g oups a eling oge he . The sys em akes in o accoun he p e e ences
o each indi idual in e ms o p ice, loca ion, and ameni ies, p o iding op ions ha a e accep able
o all.
Peach
14
is a g oup ood o de ing app, whe e i o e s ecommenda ions o es au an s ha i he
collec i e as es o he g oup. I agg ega es indi idual p e e ences and die a y es ic ions o make
g oup o de ing mo e e icien and sa is ying o e e yone. As an example, company eams o de ing
lunch h ough Peach can ge es au an ecommenda ions ha i a ious die a y needs, such as
egan, glu en- ee, o popula choices wi hin he eam.
These examples illus a e how GRS a e g adually being inco po a ed in o eal-wo ld applica ions,
enhancing collabo a i e choice-making in a ious con ex s such as en e ainmen , a el, dining,
and social planning.
2.3.2
Agg ega ion S a egies
As p e iously men ioned, o know how o combine he use s’ indi idual p e e ences is c ucial in a
GRS. Many agg ega ion s a egies ha e been de eloped, some inspi ed by Social Choice Theo y
(Mas ho , 2015). They help balance he di e se p e e ences o indi idual g oup membe s o gen-
e a e a ecommenda ion ha is accep able o he en i e g oup. This ensu es ha he ecommen-
da ion is no o e ly biased owa ds one membe ’s p e e ences (Fel e nig e al., 2023). Agg ega ion
s a egies should aim o make ai ecommenda ions by conside ing he p e e ences o all g oup
membe s. This helps a oid si ua ions whe e he majo i y’s p e e ences o e shadow hose o he
mino i y (Dueñas-Le ín, La a-Cab e a, O ega, & Bobadilla, 2023). By agg ega ing p e e ences, he
sys em can s eamline he ecommenda ion p ocess, making i mo e e icien and less ime-
consuming (Lin, Zhang, Yang, Song, & Peng, 2021). This is pa icula ly impo an when dealing
13
h ps://www.businessinside .com/google-maps-g oup-planning-polling- ea u e-how- o-use-2018-9
14
h ps://www.peachd.com
Chap e 2 – Main Concep s
32
wi h la ge g oups. E ec i e agg ega ion s a egies can adap o changes in g oup membe s’ p e -
e ences o e ime, ensu ing ha he ecommenda ions emain ele an and sa is ac o y. P ope ly
agg ega ed ecommenda ions can enhance o e all g oup sa is ac ion by ensu ing ha he selec ed
i ems o ac i i ies a e enjoyable o e e yone in ol ed (Mas ho , 2015).
Mas ho (2015) explains ele en agg ega ion s a egies, ca ego ized by Seno e al. (2010) in o
h ee majo ypes: (1) majo i y-based s a egies, like plu ali y o ing, which use he mos p e alen
i ems; (2) consensus-based s a egies, like a e age o ai ness, which ake in o accoun all g oup
membe s’ p e e ences; and (3) bo de line s a egies, like dic a o ship o leas mise y, which only
conside a subg oup. Table 2.1 summa izes he ele en agg ega ion s a egies (mo e de ails in
Mas ho (2015)).
Table 2.1. Agg ega ion s a egies used in GRS, adap ed om Mas ho (2015). Rep oduced wi h pe mission om Sp inge
Na u e.
Agg ega ion S a egy
Desc ip ion
App o al o ing
Coun s he indi iduals wi h a ings o he i em abo e an app o al h eshold
(e.g. 6)
A e age
A e ages indi idual a ings
A e age wi hou mise y
A e ages indi idual a ings, a e excluding i ems wi h indi idual a ings below
a ce ain h eshold (say 4)
Bo da coun
Coun s poin s om i ems’ ankings in he indi iduals’ p e e ence lis s, wi h
bo om i em ge ing 0 poin s, nex one up ge ing one poin , e c.
Copeland ule
Coun s how o en an i em bea s o he i ems (using majo i y o e
a
) minus how
o en i loses
Fai ness
I ems a e anked as i indi iduals a e choosing hem in u n
Leas mise y
Takes he minimum o indi idual a ings
Mos pleasu e
Takes he maximum o indi idual a ings
Mos espec ed pe son (Dic a o ship)
Uses he a ing o he mos espec ed indi idual
Mul iplica i e
Mul iplies indi idual a ings
Plu ali y o ing
Uses ‘ i s pas he pos ’: epe i i ely, he i em wi h he mos o es is chosen
a I mos g oup membe s a e i em X highe han i em Y, hen i em X is p e e ed o e i em Y.
2.4
M
ULTI
-A
GENT
S
YSTEMS AND
M
ULTI
-A
GENT
M
ICROSERVICES
Due o he complexi y associa ed o g oups o ou is s, se e al s a egies a e being p oposed o
imp o e ecommenda ions, like he use o Mul i-Agen Sys ems (Bo às e al., 2014; Ra i,
De a ajan, Sangaiah, Wang, & Sub amaniyaswamy, 2021; Z. Wang, Yu, Zheng, Ma, & Zhang,
2024). These sys ems a e composed o mul iple in e ac ing agen s, each one capable o au ono-
mous ac ion o achie e speci ic goals, being able o communica e, coo dina e, and collabo a e o
sol e complex p oblems. By using MAS, each agen can be op imized o speci ic asks, leading o
Chap e 2 – Main Concep s
33
g ea e e iciency and pe o mance (Van de Hoek & Woold idge, 2008; Woold idge & Jennings,
1995). MAS can easily adap o changing en i onmen s by adding, emo ing, o modi ying agen s
wi hou needing o e ain he en i e sys em, making hem sui able o a wide ange o applica ions
like RS. The capabili y o decen alized con ol ensu es ha he sys em con inues o ope a e e en i
some componen s ail. By dis ibu ing asks among mul iple agen s, MAS can e ec i ely manage
and il e la ge amoun s o da a, p o iding mo e ele an ecommenda ions (Do i, Kanhe e, &
Ju dak, 2018; Mahmood, El-Benda y, Pla oš, Hassanien, & He ny, 2014).
MAS a e a powe ul app oach o enhancing he pe o mance and adap abili y o RS. Thei e-
ma kable simila i ies wi h mic ose ices (MS) make hem a pe ec combina ion. Mul i-Agen MS
sp ou ed om his simila i y and combine he p inciples o MAS and MS a chi ec u e o c ea e a
mo e dynamic and lexible sys em (W. Collie , O'Neill, Lillis, & O'Ha e, 2019). The MS a chi ec u al
s yle s uc u es an applica ion as a collec ion o loosely coupled se ices. Each se ice is ine-
g ained and he p o ocols a e ligh weigh , ypically HTTP/REST. As MS allow he deploymen o
modula and scalable independen se ices, wi h independen da abases, di e en p og amming
languages can be used o each one (Villamiza e al., 2015). MAMS in eg a e hese wo concep s
by deploying agen s wi hin mic ose ices, whe e each mic ose ice can hos one o mo e agen s,
exposing hei unc ionali ies h ough REST endpoin s. Wi h his app oach, agen s can in e ac wi h
each o he and wi h ex e nal sys ems mo e e icien ly h ough s anda dized communica ion p o o-
cols; and he sys em can easily scale by adding o emo ing mic ose ices wi hou a ec ing he
o e all a chi ec u e (Jagu is, Russell, & Collie , 2023). By le e aging he s eng hs o bo h MAS and
mic ose ices, MAMS can p o ide as e and mo e in elligen esponses o use eques s (Z. Liu,
Yu, Fan, & Chen, 2022). This combina ion makes MAMS a powe ul app oach o building com-
plex, adap i e, and scalable sys ems. Table 2.2 and Table 2.3 show he compa ison be ween bo h
MAS and MAMS p inciples, showing he e e ed commonali ies, which a e ho oughly explained in
W. Collie e al. (2019).
Se e al Mul i-Agen pla o ms exis (Pal, Leon, Pap zycki, & Ganzha, 2020). A e a ho ough e-
sea ch, Ac essMAS was chosen o de eloping he GRS p o o ype’s MAMS (see Pape 4, Chap e
6), o being a simple amewo k wi h an easy lea ning cu e. Addi ionally, i is one o he ew
agen amewo ks ha u ilize he C# p og amming language, making i mo e compa ible wi h he
Chap e 2 – Main Concep s
34
.NET amewo k, chosen o he p o o ype de elopmen . Ac essMAS is buil on he Ac o model
and le e ages .NET asynch onous ope a ions o i s implemen a ion. To ensu e a anspa en and
e icien pa allel and sequen ial execu ion, in Ac essMAS he “agen s a e s o ed in a concu en o
‘no mal’ C# dic iona y” (Leon, 2022). The agen s a e ali e in an en i onmen , execu ing in u ns,
and communica e wi h each o he using messages. The messaging sys em uses a simple syn ax,
based on he FIPA’s Agen Communica ion Language (ACL) s anda d and he JADE amewo k
messages implemen a ion. The pa allel execu ion can bene i om he mul iple p ocesso co es
now a ailable, and is pe o med by using .NET Tasks, which use a h ead pool ha allows he exe-
cu ion o housands o agen s a he same ime.
Table 2.2. Compa ison o he mic ose ice p inciples o MAS. Adap ed om W. Collie e al. (2019), licensed unde CC BY 4.0
(h ps://c ea i ecommons.o g/licenses/by/4.0/).
P inciple
Mic ose ices
MAS
Bounded Con ex
A mic ose ice ep esen s a single piece o
business unc ionali y.
An agen can play a single o mul iple oles
in a sys em.
Size
Mic ose ices should be small enough o
ensu e main ainabili y
and ex ensibili y.
Size/complexi y is no an issue in MAS
esea ch and o en depends
on he a ge
domain.
Isola ed S a e
Sha ing o s a e in o ma ion is minimized
ac oss se ices.
S a e is local and p i a e o an agen . This
is o en iewed as essen ial o an agen s
au onomy
.
Dis ibu ion
Se ices a e sp ead
ac oss mul iple nodes.
Agen s a e logically dis ibu ed, bu i is
also expec ed ha hey will be sp ead o e
mul iple nodes.
Elas ici y
The applica ion is designed o allow
addi ion and emo al o equi ed esou ces
a
un ime.
The abili y o add/ emo e agen s a
un ime is a cen al ea u e
o MAS.
Au oma ed
Managemen
Managemen ope a ions like ailu e
handling and scaling
a e au oma ed.
Managemen ope a ions a e no cen al o
agen s, bu a e
some imes conside ed.
Loose
Coupling
Sys ems a e decomposed in o loosely
coupled se s o highly-cohesi e
coloca ed
se ices.
Agen s a e au onomous and loosely-
coupled
p oblem sol e s.
2.5
P
ERSONALITY
Pe sonali y e e s o he unique pa e ns o hough s, eelings, and beha io s ha dis inguish an
indi idual om o he s. As de ined by Allpo (1961) “Pe sonali y is he dynamic o ganiza ion wi hin
he indi idual o hose psychophysical sys ems ha de e mine his cha ac e is ics beha io and
hough ”. They a e seen as “ he cha ac e is ics o blend o cha ac e is ics ha make a pe son
unique” (Weinbe g & Gould, 2023), and hey a e conside ed s able du ing mos o an indi idual’s
li e ime independen ly o he si ua ion (McC ae & Cos a J , 1997).
Chap e 2 – Main Concep s
41
adi ional games, especially in an e a whe e he new gene a ion is su ounded by echnology and
easily ge s bo ed by he adi ional (Tsionas & Sa a zemi, 2023). While hey inco po a e elemen s
o adi ional gameplay, such as ules, objec i es, challenges, and ewa ds, hei p ima y ocus is
on educa ion, aining, o sol ing eal-wo ld p oblems. These games a e being success ully used in
ields like educa ion (e.g., “Minec a : Educa ion Edi ion”
17
(Ba -El & E. Ringland, 2020) and “Ci i-
liza ion” (Mol, Poli opoulos, & A iese-Vandemeuleb oucke, 2017)), heal hca e (e.g., Ad en u e
Game (Ven u ini, Fa alla, & Innocen i, 2013) and Foldi (Coope e al., 2010)), business, mili a y
(e.g., Ame ica’s A my (Niebo g, 2004)), and public policy, o engage use s in lea ning o skill-
building h ough in e ac i e expe iences.
As men ioned be o e, in his wo k, one in en ion was o implici ly de e mine he use s’ pe sonali y.
Se e al wo ks ha y o unob usi ely de e mine he use s’ pe sonali y by using se ious games a e
al eady being de eloped (see Pape 7, Chap e 9). Fo ins ance, he s o y-based se ious game
“Lea n C”, de eloped by Tsionas and Sa a zemi (2023), was de eloped o help s uden s lea n he
C p og amming language, and a he same ime, o model he s uden s’ pe sonali y. They ound
ex a e sion and openness could be e ec i ely measu ed wi h he game. Howe e , o he bes o
ou knowledge, he exis ing wo ks s ill need many use s’ in e ac ions and inpu o cha ac e ize
hem, ake oo much ime o be played, a e no ha accu a e, o only de e mine he b oade i e
dimensions o pe sonali y, ins ead o he mo e g anula 30 ai s. So, he challenge is o de elop
an imme si e and appealing game ha can lea n mo e g anula pe sonali y ai s in a sho ime.
To implici ly acqui e someone’s pe sonali y poses o he challenges, such as he p i acy, con iden-
iali y, and anonymi y o he use ’s pe sonal da a, which mus be conside ed and assu ed in he
design o his ype o games, pa icula ly in ela ion o sensi i e da a such as hei pe sonali y, as
manda o y by he Gene al Da a P o ec ion Regula ion (GDPR) in Eu ope.
17
h ps://educa ion.minec a .ne /en-us/ge -s a ed/school-leade s

Pa II
THE PAPERS
43
“Mis akes a e he building blocks o wisdom”
Ma ha Rodge s, in Cas le
44
3
M O D E L I N G A M O B I L E G R O U P R E C O M M E N D E R
S Y S T E M F O R T O U R I S M W I T H I N T E L L I G E N T
A G E N T S A N D G A M I F I C A T I O N
An image ep esen ing a concep ual GRS whe e he use s a e ep esen ed by agen s.
Gene a ed wi h AI, Image C ea o in Bing, Oc obe 2024
45
Pape Ti le
Modeling a Mobile G oup Recommende Sys em o Tou ism
wi h In elligen Agen s and Gami ica ion
Au ho s
Pa ícia Al es, João Ca nei o, Go e i Ma ei os, Paulo No ais
Publica ion Type
Con e ence P oceedings
Con e ence Name
HAIS 2019: The 14 h In e na ional Con e ence on Hyb id A i icial In elli-
gence Sys ems
Yea
2019
Place
León, Spain
DOI
h ps://doi.o g/10.1007/978-3-030-29859-3_49
ISBN
978-3-030-29858-6
Published
26 Augus 2019
Publishe
Sp inge Na u e, Cham
URL
h ps://link.sp inge .com/chap e /10.1007/978-3-030-29859-3_49
CORE2018
Rank C
Chap e 3 (Pape 1) - Modeling a Mobile G oup Recommende Sys em o Tou ism wi h In elligen Agen s and Gami ica ion
46
Abs ac
To p o ide ecommenda ions o g oups o people is a complex ask, especially due o he g oup’s
he e ogenei y and con lic ing p e e ences and pe sonali ies. This he e ogenei y is e en deepe in
occasional g oups o med o p ede ined ou packages in ou ism. G oup Recommende Sys ems
(GRS) a e being designed o helping in si ua ions like hose. Howe e , many limi a ions can s ill be
ound, ei he on hei ime-consuming con igu a ions and excessi e in usi eness o build he ou -
is s’ p o ile, o in hei lack o conce n o he ou is s’ in e es s du ing he planning and ou s, like
eeling a g ea e libe y, diminish he sense o ea /being los , inc ease hei sense o companion-
ship, and p omo e he social in e ac ion among hem wi hou losing a pe sonalized expe ience. In
his pape , we p opose a concep ual model ha in ends o enhance GRS o ou ism by using gam-
i ica ion echniques, in elligen agen s modeled wi h he ou is s’ con ex and p o ile, such as psy-
chological and socio-cul u al aspec s, and dialogue games be ween he agen s o he pos -
ecommenda ion p ocess. Some impo an aspec s o a GRS o ou ism a e also discussed, open-
ing he way o he p oposed concep ual model, which we belie e will help o sol e he iden i ied
limi a ions.
Keywo ds
: G oup Recommende Sys ems, Mobile Tou ism, Con ex -Awa eness, Gami ica ion,
Mul i-Agen Sys ems
3.1
I
NTRODUCTION
Since 1992 (D. Goldbe g e al., 1992) ha Recommende Sys ems (RS) ha e been s udied o help
indi idual use s make be e choices (Jameson e al., 2015; Resnick & Va ian, 1997) hus ecom-
mending i ems ha in end o be e sa is y he use s as es in a ious domains, each one wi h i s
speci ic challenges, like ecommending a mo ie o wa ch, a music o lis en, a place o isi , a es-
au an o lunch, e c. Bu i o gene a e accu a e indi idual ecommenda ions is complex, o p o-
ide accu a e ecommenda ions o g oups is e en mo e. The ou ism domain has many pa icula i-
ies and is an in e es ing challenge. To suppo g oups o ou is s plan and ge accompanied in
hei excu sions can be a e y complex ask, especially due o he g oup’s he e ogenei y and con-
lic ing p e e ences (T. N. Nguyen & Ricci, 2018). Millions o ou is s pa icipa e in planned ou s
e e y day, some a el alone, o he s in g oups, bu a e hei needs, in e es s and cu iosi y sa is-

Chap e 3 (Pape 1) - Modeling a Mobile G oup Recommende Sys em o Tou ism wi h In elligen Agen s and Gami ica ion
47
ied? Do hey enjoy he ou s hey engaged in? Bo a o and Ca a (2010) s a e how a g oup is
o med in luences i s modeling and he p edic ed ecommenda ions. G oups o med occasionally
o a common aim, like a elling oge he o a speci ic des ina ion, and ha may o may no be
acquain ed o each o he (Bo a o & Ca a, 2010) causes his he e ogenei y o go deepe . G oup
Recommende Sys ems (GRS) a e being designed o helping in si ua ions like hose, and i hey
use he capabili ies o a mobile de ice, hey can b u ally imp o e he use s’ expe ience, b inging
new possibili ies o explo e, like he use s’ con ex (del Ca men Rod íguez-He nández e al., 2017),
i.e., he in o ma ion ha su ounds him (Lams us, Wang, Alzua-So zabal, & Xiang, 2015).
In his pape , we in oduce a concep ual model ha in ends o imp o e he ou is s expe ience in a
GRS o ou ism by showing conce n o hei in e es s, acili a e he pos - ecommenda ion p o-
cess, by p oposing he use o an a gumen a ion-based dialogue model be ween in elligen agen s,
agen s ha will accompany he ou is s du ing he ou . Gami ica ion echniques a e also p oposed
o acqui e he ou is s’ p o ile and mo i a e hem du ing he ou .
In he nex sec ion we p esen a b ie s a e-o - he-a in GRS o ou ism and discuss some cu en
issues. Sec ion 3 in oduces dialogue games be ween in elligen agen s and gami ica ion as ways
o enhancing he choice p ocess and he ou is s’ in ol emen in GRS o ou ism, espec i ely.
This sec ion also explains he connec ion be ween choice and decision, and how impo an expla-
na ions a e in a ecommenda ion. The concep ual model o he GRS o ou ism is p esen ed and
sho ly explained. Sec ion 4 summa izes he con en s add essed in he pape and desc ibes wha
will be done as u u e wo k.
3.2
G
ROUP
R
ECOMMENDER
S
YSTEMS FOR
T
OURISM
GRS ha e become an impo an and challenging heme in he ield o RS (Cas o e al., 2015; Delic
& Mas ho , 2018; Mas ho , 2011, 2015; McCa hy, Salamó, e al., 2006b) since he g oup
membe s’ p e e ences can a y, and he e o e, o each a solu ion ha sa is y all he membe s
can be ha d o accomplish. I is o ex eme impo ance o gua an ee ha none o he g oup mem-
be s ge s oo dissa is ied, dissa is ac ion ha can sp ead wi hin he g oup due o he emo ional
con agion phenomenon (Delic & Mas ho , 2018). Fo ins ance, suppose a a el agency in China
ha has aca ion packages o g oups o ou is s, wi h a se o di e en ypes o Poin s o In e es
Chap e 3 (Pape 1) - Modeling a Mobile G oup Recommende Sys em o Tou ism wi h In elligen Agen s and Gami ica ion
48
(POI) o isi in a ce ain coun y. I is known ha Chinese ou is s usually a el in g oups, ei he
by op ion o because o imposi ions (Nasolomampionona, 2014). Families, indi iduals, iends can
subsc ibe a package. Bu does he package has POI ha sa is y all he subsc ibed membe s? Al -
hough hey sha e he same cul u e, no all membe s ha e he same pe sonali y and p e e ences,
bu hey had no o he choice han o choose a p ede ined package. A aca ion ha seemed exci -
ing can easily become oilsome. A GRS capable o p o iding pe sonal and con ex ual ecommenda-
ions can be he pe ec solu ion.
Many in e es ing p o o ypes o GRS o ou ism ha e and a e being p oposed o help g oups o
ou is s in he planning o aca ions o excu sions, usually p esen ing a lis o POI o isi . Fo in-
s ance, looking a some o he i s GRS o ou ism,
INTRIGUE
(IN e ac i e TouRis In o ma ion
GUidE) was p oposed in 2003 by A dissono e al. (A dissono e al., 2003) o help (he e ogenous)
g oups o ou is s ind sigh seeing des ina ions and i ine a ies in I aly. I is a GRS o mobile and
desk op de ices whe e a g oup membe con igu es he g oup size, hei p e e ences and cha ac-
e is ics. The g oup is hen di ided in o subg oups acco ding o hose con igu a ions, and ecom-
menda ions a e gi en o each subg oup g ounded by explana ions ha add ess po en ial con lic ing
equi emen s.
CATS
(Collabo a i e Ad iso y T a el Sys em) aims o help a g oup o iends in planning a ski-
holiday (McCa hy, McGin y, Smy h, & Salamó, 2006; McCa hy, Salamó, e al., 2006b) using a
ace- o- ace collabo a i e pla o m ( he DiamondTouch in e ac i e able op) ha uses c i iques as a
way o gi ing eedback o ecommended POI and i e a i ely ind a inal choice.
Ga cia e al. (Ga cia e al., 2009) de eloped a GRS o ou is ac i i ies, based on he g oup’s
as es, demog aphic da a and places isi ed in o me ips, by ex ending he
e-Tou ism
ool hey
p e iously de eloped o indi idual ou is s. This ool is composed by he
Gene alis Recommende
Sys em Ke nel
(GRSK), which is a domain-independen axonomy-d i en sea ch engine ha man-
ages he g oup ecommenda ion. I is esponsible o agg ega ing, in e sec ion and inc emen ally
in e sec ion he use s’ p e e ences and p esen a inal lis o i ems o ecommend.
T a el Decision Fo um is a GRS ha uses anima ed cha ac e s o ep esen he g oup membe s
(Jameson e al., 2003). The au ho s s a e ha mu ual-awa eness and communica ion a e im-
po an in o de o each a consensus in he pos - ecommenda ion p ocess. Fo ha , he g oup
Chap e 3 (Pape 1) - Modeling a Mobile G oup Recommende Sys em o Tou ism wi h In elligen Agen s and Gami ica ion
49
membe s con igu e hei p e e ences inc emen ally and collabo a i ely, being able o see he o he
membe s’ p e e ences. Since he choice o p e e ences can be in luenced by a pe son’s mo i a-
ions, he au ho s implemen ed a simple way o he membe s o con igu e hei mo i a ional o i-
en a ion ega ding he o he membe s. This is a e y impo an ac o in social in e ac ions ha
o he GRS do no conside , and ha we will u he discuss la e in his pape .
I is pe cep ible ha due o echnological limi a ions a he ime, he i s GRS we e o ally depend-
en o he use s’ in e ac ions and con igu a ions. Indeed, since he mobile echnology was s ill
eme ging, he use s el o ended o ha ing a “ oo in elligen ” applica ion and a gued hey could
hink and decide o hemsel es, no accep ing a oo much au oma iza ion o he sys em (Van
Se en, Pok ae , & Koolwaaij, 2004). Howe e , i een yea s la e , he minds “e ol ed”, he use s’
equi emen s changed, and many would like o ha e a mo e au oma ed sys em ha could hink
and decide o hem, a leas ega ding ecommenda ions...
In he ea ly 2000’s, wi eless in e ne access was e y limi ed and e y expensi e, bu now, ha is
no longe a p oblem. The apid e olu ion o he wi eless in e ne connec ions, i s h oughpu , s abil-
i y, p ice and massi ica ion, also shi ed he way (G)RS we e being designed and many ide-
as/app oaches ound in li e a u e we e discon inued. This is a posi i e ein o cemen o c ea ing
new and be e (G)RS.
Fo example, he e y ecen wo k by T. N. Nguyen and Ricci (2018) consis s on a cha -based GRS
o mobile de ices ha also allows he g oup membe s o become pa o he choice p ocess. I is
simila o Wha sApp in he way use s in a g oup can exchange messages be ween hem, wi h he
addi ional ea u es o allowing he use s o a e p e iously isi ed POI and de ine hei mood, so a
highe impo ance is a ibu ed o he use in he p e e ences agg ega ion in case he is in a bad
mood, i ed, e c. The use s can classi y he ecommended POI by liking/disliking hem o by classi-
ying one as he bes , o commen on hem wi h ex and emo icons. This e alua ion allows he
sys em o in e use s’ cons ain s based on he a ibu es o he classi ied POI, and inc emen ally
upda e he in o ma ion on a ecommended POI wi h addi ional explana ions, based on hose e-
s ic ions. Al hough he sys em p o ided highe pe cei ed ecommenda ion quali y han he s and-
a d benchma k, his app oach may no be p ac ical o la ge and/o occasional g oups, since he
es ed g oups we e e y small, composed o 2 o 3 membe s. We hink i can be e y con using o
Chap e 3 (Pape 1) - Modeling a Mobile G oup Recommende Sys em o Tou ism wi h In elligen Agen s and Gami ica ion
50
a g oup o 20 o mo e people o cha and exchange opinions in an e icien way. Some hing else is
needed.
3.2.1
Impo an Aspec s o Conside in a GRS o Tou ism
To suppo g oups in a el planning is no a simple p ocess and o gene a e a lis o ecommenda-
ions based on he use s’ con ex and p e e ences is no enough. O he ac o s need o be consid-
e ed o a GRS o e ec i ely se e i s pu poses. Fo ins ance, in 2003, Jameson e al. (2003)
made he in elligen obse a ion ha he ecommenda ion p ocess does no end when a lis o
ecommenda ions is p esen ed o he use . The use s need o decide wha o choose om he lis ,
so all he g oup membe s ge (minimally) sa is ied. The au ho s wen e en u he by s a ing ha i
would be sho -sigh ed no o include pos - ecommenda ion p ocesses in he design o a (G)RS, like
ways o pe suading he o he g oup membe s o ollow a ce ain ecommenda ion a use inds
be e . I he p ocess o eaching he inal choice has no been delega ed o one o he g oup
membe s, communica ion and possibly nego ia ion will be needed be ween he g oup membe s
(Jameson e al., 2003). This alls in o he same line o hough ha he use s need o be somehow
in ol ed in he ecommenda ion p ocess, and as men ioned be o e, a ull au oma iza ion may no
be he pe ec solu ion.
I is e idenced ha many people like o know he p e e ences o o he g oup membe s, leaning o
choose simila p e e ences (Jameson e al., 2003), ei he because hey would like o please o he
membe (s) o because hey end o a oid con lic s i hey p e iously know wha he o he use s
hink, like in a eal ace- o- ace scena io. This awa eness leads o a so o collabo a ion ha can
help each a as e consensus. Howe e , his ype o beha io is no so linea . Like in a decision-
making p ocess, he g oup membe s in a choice p ocess can ha e di e en in en ions, which in-
luence hei beha io s and choices. Jameson e al. (2003) add ess mo i a ion as a way o in lu-
encing he choice p ocess. Howe e , mo i a ion is wha compels us o ul il o no ou in en ions.
So, a pe son’s in en ions a e in he co e o a choice, powe ed by he mo i a ions, and we belie e
bo h need o be accoun ed o . Fo ins ance, Phoebe can ha e an in en ion o isi a coun y, bu
because she canno go wi h he boy iend, she doesn’ eel mo i a ed o go, and he e o e she
won’ go unless he does.
Chap e 3 (Pape 1) - Modeling a Mobile G oup Recommende Sys em o Tou ism wi h In elligen Agen s and Gami ica ion
57
ou is ’s in en ions and in e es s. Why an a a a ? I is e idenced ha ep esen ing he ou is wi h
an a a a can help him eel empa hy owa ds he sys em (Mo a a e al., 2014).
Loca ion-based AR games can ha e a emendous po en ial, and hey can be a sma e way o
ca ching he ou is s’ a en ion o isi a coun y’s he i age. We p opose o ans o m he whole ip
p ocess in o a so o a loca ion-based AR game, whe e he ou is s will ha e o comple e ce ain
pe sonalized “ques s” in he POI hey isi , using AR ea u es. We hope his will also inc ease hei
in e es in knowing and lea ning abou a coun y’s he i age, and in a mo e exci ing way.
3.4
S
UMMARY AND
F
UTURE
W
ORK
In his wo k, we discuss on a no el app oach o a G oup Recommende Sys em o ou ism using
agen s and gami ica ion. The aim is no o ocus on a be e algo i hm o gene a ing a lis o ec-
ommenda ions, bu o acili a e he consensus in he pos - ecommenda ion p ocess so highe qual-
i y and mo e sa is ac o y choices can be made, and o enhance he ou is s’ expe ience du ing he
whole p ocess, om he planning o he ou i sel . We in end o accomplish his by aking ad-
an age o dialogue games using a gumen a ion o he pos - ecommenda ion p ocess, be ween
in elligen agen s modeled wi h he ou is s’ p o ile and con ex , and by in oducing gaming com-
ponen s in he sys em ha will encou age he ou is s’ in e ac ion in a mo e appealing way. The
ou is s’ p o ile and con ex will be used o p o ide mo e in elligen and pe sonalized ecommenda-
ions and no i ica ions du ing he whole ou , o g oups o any size. We belie e he dialogue games
be ween he agen s will be a sma e way o explaining he ecommenda ions o he ou is s.
T a elling is an emo ional expe ience (Delic, Neidha d , Nguyen, & Ricci, 2016) and he e o e,
pe sonaliza ion and gami ica ion a e becoming a c ucial ac o o he success o GRS in ou ism.
In ac , gami ica ion echniques and pe sonalized se ices will be a majo end o he u u e o
ou ism (Xu e al., 2016). To mo i a e he ou is s in planning he g oup ou and con igu e hei
p o ile and con ex , ei he implici ly o explici ly, we p opose he use o gami ica ion echniques like
mini games, badges, ophies, and ankings o he bes achie emen s. An AR a a a is also p o-
posed o ep esen he ou is ’s agen and accompany him h ough he whole p ocess, including
du ing he ou , being esponsible o p o iding pe sonalized and con ex ual ecommenda ions and
push-no i ica ions o he ou is ’s well-being.

Chap e 3 (Pape 1) - Modeling a Mobile G oup Recommende Sys em o Tou ism wi h In elligen Agen s and Gami ica ion
58
The p oposed app oach will be ho oughly explained in ou u u e wo k, and will include, among
o he asks, he ealiza ion o ques ionnai es o di e en cul u es in o de o de elop he model o
co ela e pe sonali y ai s wi h (cul u e ela ed) ou is ic p e e ences, and he de elopmen o mini
games o implici ly acqui e he ou is s’ pe sonali y, p e e ences and con ex . The ga he ed in o -
ma ion will be used o model he agen s ep esen ing he ou is s and hei a a a . The Social Ne -
wo k p o o ype will be de eloped o he pos - ecommenda ion choice p ocess and o enable he
ou is s’ online in e ac ion. In elligen push-no i ica ions, ecommenda ions, o he mini games and
asks du ing he ou will be designed based on he ou is s’ p o ile and con ex . Expe imen s wi h
eal use s will be conduc ed o es he iabili y o he p oposed wo k and he use s’ sa is ac ion.
Acknowledgemen s
This wo k was suppo ed by he G ouPlanne P ojec (POCI-01-0145-FEDER-29178) and by Na-
ional Funds h ough he FCT – Fundação pa a a Ciência e a Tecnologia (Po uguese Founda ion
o Science and Technology) wi hin he P ojec s UID/CEC/00319/2019 and UID/EEA/00760/
2019.
59
4
M O D E L I N G T O U R I S T S ' P E R S O N A L I T Y I N R E C O M -
M E N D E R S Y S T E M S : H O W D O E S P E R S O N A L I T Y
I N F L U E N C E P R E F E R E N C E S F O R T O U R I S T A T -
T R A C T I O N S ?
An image ep esen ing dis inc ou is a ac ions and he ou is s’ di e en pe sonali ies.
Gene a ed wi h AI, Image C ea o in Bing, Oc obe 2024
Chap e 4 (Pape 2) - Modeling Tou is s' Pe sonali y in Recommende Sys ems: How Does Pe sonali y In luence…
60
Pape Ti le
Modeling Tou is s' Pe sonali y in Recommende Sys ems: How
Does Pe sonali y In luence P e e ences o Tou is A ac ions?
Au ho s
Pa ícia Al es, Ped o Sa ai a, João Ca nei o, Ped o Campos, Helena
Ma ins, Paulo No ais, Go e i Ma ei os
Pape Type
Con e ence P oceedings
Con e ence Name
UMAP '20: 28 h ACM Con e ence on Use Modeling, Adap a ion and
Pe sonaliza ion
Yea
2020
Place
Genoa, I aly (online)
DOI
h ps://doi.o g/10.1145/3340631.3394843
ISBN
9781450368612
Published
13 July 2020
Publishe
Associa ion o Compu ing Machine y (ACM)
URL
h ps://dl.acm.o g/doi/abs/10.1145/3340631.3394843
CORE2020
Rank B
Chap e 4 (Pape 2) - Modeling Tou is s' Pe sonali y in Recommende Sys ems: How Does Pe sonali y In luence…
61
Abs ac
Pe sonaliza ion is inc easingly being pe cei ed as an impo an ac o o he e ec i eness o Rec-
ommende Sys ems (RS). This is especially ue in he ou ism domain, whe e a elling comp ises
emo ionally cha ged expe iences, and he e o e, he mo e abou he ou is is known, be e ec-
ommenda ions can be made. The inclusion o psychological aspec s o gene a e ecommenda-
ions, such as pe sonali y, is a g owing end in RS and hey a e being s udied o p o ide mo e
pe sonalized app oaches. Howe e , al hough many s udies on he psychology o ou ism exis ,
s udies on he p edic ion o ou is p e e ences based on hei pe sonali y a e limi ed. The e o e,
we unde ook a la ge-scale s udy in o de o de e mine how he Big Fi e pe sonali y dimensions
in luence ou is s’ p e e ences o ou is a ac ions, ga he ing da a om an online ques ionnai e,
sen o Po uguese indi iduals om he academic sec o and hei espec i e ela i es/ iends
(n=508). Using Explo a o y and Con i ma o y Fac o Analysis, we ex ac ed 11 main ca ego ies o
ou is a ac ions and analyzed which pe sonali y dimensions we e p edic o s (o no ) o p e e -
ences o hose ou is a ac ions. As a esul , we p opose he i s model ha ela es he i e pe -
sonali y dimensions wi h p e e ences o ou is a ac ions, which in ends o o e a base o e-
sea che s o RS o ou ism o au oma ically model ou is p e e ences based on hei pe sonali y.
CCS Concep s:
In o ma ion sys ems ~ Recommende sys ems • Human-cen e ed compu ing ~
Use cen e ed design
Keywo ds:
Recommende Sys ems; Pe sonali y; Tou is P e e ences; A ec i e Compu ing; Lei-
su e Tou ism
4.1
I
NTRODUCTION
Recommende Sys ems (RS) a e being s udied in a ious domains o help use s make be e
choices (Adoma icius & Tuzhilin, 2005; Resnick & Va ian, 1997), being one widely s udied he
a el and ou ism domain. T a elling is an emo ional expe ience (Delic, Neidha d , Nguyen, &
Ricci, 2016) and he e o e, pe sonaliza ion is a key ac o o he success o RS in ou ism (Ga alas
& Ken e is, 2011; Ricci, 2002; Schmid -Belz e al., 2002). The mo e in o ma ion abou he ou is
is known be e ecommenda ions can be made. A i icial In elligence echniques, like Mul i-Agen
Sys ems, ha e been applied o RS o y o enhance he ou is s’ expe ience and p oac i ely make
Chap e 4 (Pape 2) - Modeling Tou is s' Pe sonali y in Recommende Sys ems: How Does Pe sonali y In luence…
62
sugges ions based on he ou is s’ con ex and p o ile (Ba e , Mo eno, Sánchez, Ise n, & Valls,
2012; Bo às e al., 2014; Cecca oni, Codina, Palau, & Pous, 2009; Lo enzi, Loh, & Abel, 2011).
To pe sonalize agen s wi h he ou is s’ p o ile (Mo a a e al., 2014), o accompany hem h ough-
ou he p ocess by p esen ing in elligen in o ma ion and p oposing pe sonalized challenges, ac-
co ding o he ou is s’ psychological aspec s and in e es s, can imp o e hei expe ience and sa is-
ac ion. Fo ins ance, pe sonali y has been shown o imp o e (g oup) ecommenda ions and can
e en help wi h he cold-s a p oblem (Feil e al., 2016; Tkalcic & Chen, 2015). S udies show pe -
sonali y is s ongly ela ed o he use s’ p e e ences (Tkalcic & Chen, 2015), and in he case o
ecommenda ions o g oups, co ela ing he use s’ pe sonali ies and hei p e e ences can help
ma ch use s wi h simila in e es s, minimizing he g oups’ he e ogenei y and con lic s o in e es in
(occasional) g oups o ou is s. Se e al s udies exis on he ela ion be ween pe sonali y and ou is
p e e ences, howe e , he ones a ailable only ocus on speci ic ypes o a elling o ou is oles
(Delic, Neidha d , & We hne , 2016; Eachus, 2004; Jani, 2014b; Masie o, Qiu, & Zol an, 2019; K.
Y. Poon & Huang, 2017; Schneide & Vog , 2012), o mainly a ge he Ex a e sion and Openness
o Expe ience pe sonali y dimensions (Bujisic, Bilgihan, & Smi h, 2015; C.-Y. Li, Lu, Tsai, & Yu,
2015). So, wha combina ion (i any) o pe sonali y dimensions in luence he choice o ce ain ou -
is a ac ions?
In o de o s a o e coming hose limi a ions, we engaged on a la ge-scale s udy o de e mine he
ela ion be ween he Big Fi e pe sonali y dimensions and p e e ences o ou is a ac ions. Fi s ,
an in ensi e esea ch was conduc ed so ha a ques ionnai e o collec as much in o ma ion as
possible abou he esponden s’ pe sonali y, ou is and pe sonal p e e ences, a el mo i a ions
and socio-demog aphy could be cons uc ed. The ques ionnai e was sen in a i s ound o Po u-
guese indi iduals om he academic sec o and hei espec i e ela i es/ iends, ob aining a
sample o 508 alid esponses. A e analyzing and ea ing he esponses, all pe sonali y dimen-
sions we e ound o be p edic o s o di e en p e e ences o ou is a ac ions. As a esul , we
p opose a model ha ela es all i e pe sonali y dimensions wi h p e e ences o a wide ange o
ou is a ac ions, in he hope we can help esea che s o RS o ou ism o au oma ically model
ou is p e e ences based on he ou is s pe sonali y.

Chap e 4 (Pape 2) - Modeling Tou is s' Pe sonali y in Recommende Sys ems: How Does Pe sonali y In luence…
63
The emainde o he pape is s uc u ed as ollows: Sec ion 2 desc ibes some ela ed wo k on
pe sonalized Recommende Sys ems and Psychology o Tou ism. Sec ion 3 p esen s he me hod-
ology used, Sec ion 4 he esul s and hei espec i e analysis, along wi h he p oposed models
ha ela e P e e ences o Tou is A ac ions and how pe sonali y ai s in luence hose p e e -
ences, and inally, Sec ion 5 e lec s on he con en s add essed in he pape and desc ibes wha
will be done as u u e wo k.
4.2
B
ACKGROUND
4.2.1
Pe sonali y and Recommende Sys ems
Th oughou he las wo decades, pe sonaliza ion became he main conce n o he e ec i eness o
RS (L. Chen e al., 2016; Rashid e al., 2002; Tkalcic & Chen, 2015; Tondello e al., 2017). The e-
o e, o know pe sonal in o ma ion abou he use is c ucial o building a obus p o ile. The e-
sea che s inc eased in e es in de eloping mo e pe sonalized and accu a e RS c ea ed he need o
conside o he esea ch a eas such as Psychology. Psychological aspec s, such as pe sonali y,
moods and emo ions, a e being pe cei ed o in luence he a iance in he use p e e ences and
beha io in RS (Tkalčič, De Ca olis, de Gemmis, Odić, & Koši , 2016), and hei conside a ion is
e idencing o show be e esul s han gene ic app oaches (G e zel, Mi sche, Hwang, &
Fesenmaie , 2004; Nunes e al., 2008; Tondello e al., 2017). Many pe sonalized RS ha ake in o
accoun he use s mo i a ions (Jameson e al., 2003), mood (del Ca men Rod íguez-He nández e
al., 2017; T. N. Nguyen & Ricci, 2018), o pe sonali y (Hu & Pu, 2009; Pe ik e al., 2004;
Roshchina, 2012; Tkalcic e al., 2009; Wecke , Ku lik, & S ock, 2016) ha e been de eloped.
As de ined by H. Eysenck and Rein (1998), “pe sonali y is he sum- o al o he ac ual o po en ial
beha io -pa e ns o he o ganism, as de e mined by he edi y and en i onmen ”. Each indi idual
has he own beha io pa e ns, which a e conside ed ela i ely s able o e ime ac oss di e en
si ua ions (McC ae & Cos a J , 1997). These pa e ns we e summa ized in o i e uni e sal pe son-
ali y dimensions by Cos a and MacC ae (1992): Openness o Expe ience, Conscien iousness, Ex-
a e sion, Ag eeableness, and Neu o icism, being he Fi e Fac o Model (FFM), o Big Fi e, ecog-
nized as he mos widely accep ed model o ep esen hem (Digman, 1990; Ma z e al., 2016).
Chap e 4 (Pape 2) - Modeling Tou is s' Pe sonali y in Recommende Sys ems: How Does Pe sonali y In luence…
64
Each ac o is de ined by six ai s/ ace s (Cos a & MacC ae, 1992), esul ing in a o al o 30 ai s,
which a e mo e g anula and can be used o be e cha ac e ize a pe son (see Table 4.1).
Table 4.1. Pe sonali y dimensions and hei espec i e six ai s (adap ed om Cos a and MacC ae (1992)).
Neu o icism
Ex a e sion
Openness o expe ience
Ag eeableness
Conscien iousness
Anxie y
F iendliness
Imagina ion
T us
Sel -e icacy
Ange
G ega iousness
A is ic in e es s
Mo ali y
O de liness
Dep ession
Asse i eness
Emo ionali y
Al uism
Du i ulness
Sel -consciousness
Ac i i y le el
Ad en u ousness
Coope a ion
Achie emen -s i ing
Immode a ion
Exci emen seeking
In ellec
Modes y
Sel -discipline
Vulne abili y
Chee ulness
Libe alism
Sympa hy
Cau iousness
As no ed by Tkalcic and Chen (2015), pe sonali y can be use ul in di e en a eas o RS, since i is
s ongly ela ed o he use s’ p e e ences (Can ado & Fe nández-Tobías, 2014). Use s wi h simila
pe sonali ies end o choose simila i ems o con en s (Can ado , Fe nández-Tobías, & Bellogín,
2013). Fo example, ex a e s who a e dependen on wa m h and g ega iousness end o enjoy
popula music, and pe sons who sco e high on exci emen seeking end o enjoy ock music (Can-
ado e al., 2013; Rawlings & Cianca elli, 1997). In games, ex a e s a e mo e inclined o g oup
ac i i ies han solo ac i i ies (Yee e al., 2011). E en ce ain ea u es o Ins ag am pic u es a e
ela ed o pe sonali y ai s (Fe we da, Schedl, & Tkalcic, 2015). Pe sonali y is he e o e a powe ul
cha ac e is ic o humans ha can be used o help p edic hei p e e ences in a wide ange o do-
mains, bu i is s ill an unde explo ed opic in RS (Can ado & Fe nández-Tobías, 2014). And how
abou he ou ism domain? Is pe sonali y s ongly (o weakly) ela ed o ou is p e e ences, o only
ce ain pe sonali y dimensions a e? These a e he esea ch ques ions we p opose o answe .
4.2.2
Psychology o Tou ism
Resea ch on psychology o ou ism is apidly g owing, since i is e idenced ha psychological as-
pec s a e ela ed o he choice o speci ic des ina ions (Ja a i, 1987; Passa a o e al., 2015; S. C.
Plog, 1974). Bu which ones? Se e al esea che s ied o answe ha ques ion, some by p opos-
ing ou is ypologies based on psychological aspec s, o he s by ying o ind ela ionships among
pe sonali y ai s and ou is beha io s o p e e ences.
4.2.2.1
Tou is Typologies
E. Cohen (1972) was one o he i s esea che s o p opose a ou is ypology, composed o ou
ypes: he o ganized mass ou is (leas ad en u ous, lazy, p e e s package- ou s, is mo e o ga-
Chap e 4 (Pape 2) - Modeling Tou is s' Pe sonali y in Recommende Sys ems: How Does Pe sonali y In luence…
65
nized and p e e s amilia i y o no el y), he indi idual mass ou is (simila o o ganized mass bu
he ou is no ully p eplanned, has a ce ain con ol o e his ime and i ine a y and is no bound
o a g oup), he explo e ( ip sel -a anged, likes o mee locals and speak hei language wi hou
o ally imme sing he sel ), and he d i e (ex emely independen , has no ime schedules o i ine -
a y, li es wi h he locals, likes no el y a maximum and amilia iza ion a minimum).
Plog is ano he enown esea che who s udied he psychology o a el in ou ism (S. C. Plog,
1974). He a gued ha a el des ina ions a ac speci ic ypes o people (S. Plog, 2001) and p o-
posed wo main psychog aphic dimensions o cha ac e ize ou is s’ a el beha io : Allocen ics,
who a e mo e na u e ela ed, ad en u esome, cu ious, like o explo e he wo ld a ound hem, p ac-
ical, ou going, sel -con iden , seek o no el y and new expe iences; and Psychocen ics, who a e
sel -inhibi ed, anxious, non-ad en u esome, p e e he amilia in a el des ina ions, especially i
hey can d i e o hem, and places whe e hey can elax. The wo dimensions a e in he opposi e
ex emes o a no mally dis ibu ed con inuum, being his scale la e ex ended (S. C. Plog, 1991,
1994). Plog’s model became widely known, and many esea che s used o e en ied o im-
p o e/ex end i , some by co ela ing Plog’s wo dimensions wi h Ex a e sion (Hox e & Les e ,
1988; M. S. Jackson, Schmie e , & Whi e, 1999), ac i a ion heo ies (Nicke son & Ellis, 1991), o
sensa ion seeking, powe lessness and gene alized anxie y (G i i h & Albanese, 1996). Howe e ,
no signi ican co ela ions o ac ual ou is beha io we e ound.
M. Jackson, Whi e, and Whi e (2001) p oposed ou ypes o ou is s: he explo e , he ad en u e ,
he guided and he g oupie, combining he o hogonal scales o Allocen ics-Psychocen ics and
In o e sion-Ex a e sion, model which was la e s udied by M. S. Jackson and R. Inbaka an
(2006). As also sugges ed by Nicke son and Ellis (1991), he au ho s ound Ex a e sion and Allo-
cen ism we e independen cons uc s. The same canno be said o Openness o Expe ience and
Allocen ism, which showed o be co ela ed (M. S. Jackson & R. Inbaka an, 2006).
Howe e , he exis ence o ambigui y be ween he dimensions o bo h Plog’s and M. Jackson e al.
(2001) models, led Eachus (2004) o p opose a modi ica ion o hose ypologies so a mo e objec-
i e measu e o ou is p e e ences could be used: Ad en u ous p e e ence, Beach p e e ence,
Cul u al p e e ence, and Indulgen p e e ence. To do ha , hey used he B ie Sensa ion Seeking
Scale (BSSS) (Hoyle, S ephenson, Palmg een, Lo ch, & Donohew, 2002) o p edic hei p oposed
Chap e 4 (Pape 2) - Modeling Tou is s' Pe sonali y in Recommende Sys ems: How Does Pe sonali y In luence…
66
Holiday P e e ences Scale. They ound ha people wi h high sensa ion seeking alues end o p e-
e Ad en u ous and Beach holidays and no Indulgen holidays. No signi ican co ela ions we e
ound be ween Sensa ion Seeking and Cul u al holidays, bu olde people we e mo e likely o p e-
e Cul u al holidays han younge .
Based on he Cohen’s indi idual mass ou is ype (E. Cohen, 1972), Wickens (2002), p oposed
i e mic o- ypes o ou is s acco ding o a su ey conduc ed in Chalkidiki (G eece): he Cul u al
He i age ype, who we e mo e in e es ed in he cul u al, na u al and his o ical aspec s o he e-
gion; he Ra e ype, who we e a ac ed by sensual and hedonis ic pleasu es, p e e o spend
mo e ime a he beach and i s nigh clubs; he Shi ley Valen ine ype, who we e seeking o a o-
man ic expe ience wi h a “cha ming G eek gen leman”; he Heliola ous ype, who jus wan ed o
elax and sunba h; and he Lo d By on ype, who had he i ual o e u n e e y yea o he same
des ina ion, because hey enjoyed he amilia i y, nos algia and el like home.
In e es ingly, G e zel e al. (2004) p oposed 12 a el pe sonali ies and ound s ong co ela ions
be ween hei espec i e ac i i ies. Fo ins ance, Shopping Sha ks ype was ela ed o ou is s mo e
in e es ed in shopping, nigh li e, and dining.
4.2.2.2
Pe sonali y as P edic o o Tou is P e e ences.
Howe e , and as poin ed by se e al au ho s, he exis ing esea ch on ou ism beha io is mos ly
desc ip i e ins ead o p edic i e (M. Jackson e al., 2001; Schneide & Vog , 2012) which is a limi-
a ion ha needs o be o e come, i.e., wha pe sonali y dimensions o ai s a e p edic i e o he
ou is s’ ypologies o beha io s/p e e ences ound in li e a u e? Fo example, some esea che s
ocused on ad en u e ou ism (Addison, 1999; Milling on, Locke, & Locke, 2001), de eloping ad-
en u e ou ism ypologies such as “ha d ad en u e” and “so ad en u e” ypologies (Lipscombe,
1995). Since mos s udies ailed o de e mine he psychological an eceden s o so (e.g.: hiking,
hun ing, scuba di ing) and ha d (e.g.: climbing, ca e explo ing) ad en u e a ele s (Schneide &
Vog , 2012), Schneide and Vog (2012) applied Mowen’s (Mowen, 2000) 3M Model o Mo i a ion
and Pe sonali y o consume beha io , o explain he beha io o so and ha d ad en u e a el-
e s. They ound he in e es in cul u al expe iences, need o a ousal (exci emen seeking) and
need o ma e ial esou ces we e p edic o s o ha d ad en u e a el, and he in e es in cul u al
expe iences and compe i i eness o so ad en u e a el.
Chap e 4 (Pape 2) - Modeling Tou is s' Pe sonali y in Recommende Sys ems: How Does Pe sonali y In luence…
73
Figu e 4.2. Dis ibu ion o he i e pe sonali y dimensions sco es among he sample.
The BFI esponses o he 44 i ems we e hen used o p edic p e e ences o he p e iously p o-
posed ou is a ac ions, using S uc u al Equa ion Modeling and CFA, applying he maximum
likelihood me hod o es ima ion. This esul ed in ou p oposed model o “Pe sonali y-Tou is A -
ac ions P e e ence” as shown in Figu e 4
.3
. Al hough wi h lowe alues, he combina ion o he
indica o s e eal an accep able goodness o i (
χ
2/d = 2.486; CFI= 0.712; GFI= 0.680; PCFI=
0.676, PGFI= 0.632, RMSEA= 0.054; p[ msea≤0.05]p< 0.001), sugges ing ha he i ems p o ide
a sa is ac o y i (Ma ôco, 2010), hus con i ming he p oposed “Pe sonali y-Tou is A ac ions
P e e ence” model.
By analyzing he model, in e es ing p edic ions we e ound. Indi iduals wi h a highe p e e ence o
Ad en u e end o sco e highe on Ag eeableness bu lowe on Conscien iousness. This goes in line

Chap e 4 (Pape 2) - Modeling Tou is s' Pe sonali y in Recommende Sys ems: How Does Pe sonali y In luence…
74
wi h e idences ound in li e a u e, whe e less conscien ious people end o enjoy isky ac i i ies
(Delic, Neidha d , & We hne , 2016; Eachus, 2004; Schneide & Vog , 2012). Looking a he pe -
sonali y ai s desc ibed by Cos a J e al. (1995), his co ela ion sugges s spon aneous indi iduals
ge along easily wi h o he s.
Figu e 4.3. Simpli ied S uc u al Equa ion Model o he p oposed “Pe sonali y-Tou is A ac ions P e e ence” model. Fo
eadabili y, only he s a is ically signi ican alues a e p esen ed (* p < 0.05 (2- ailed), ** p < 0.01 (2- ailed), *** p < 0.001
(2- ailed)).
A highe p e e ence o Na u e- ela ed a ac ions is posi i ely p edic ed by highe Ag eeableness.
These indi iduals end o be mo e sympa he ic, conside a e and al uis ic, his may co espond o
mo e conce ned and a ac ed o na u e indi iduals, which may be somehow ela ed o he ecolog-
ical conce n o highly ag eeable pe sons as ound by K aso a (2015).
En e ainmen and Nigh li e ac i i ies a e posi i ely associa ed o all pe sonali y dimensions excep
Conscien iousness. One can p esume almos e e yone enjoys some o m o en e ainmen o ou -
ings, bu , once again, he e is some isk a oidance as well as a sense o ugali y ela ed o consci-
en ious pe sons ha does no seem o combine wi h u ile spending and/o unheal hy habi s.
Chap e 4 (Pape 2) - Modeling Tou is s' Pe sonali y in Recommende Sys ems: How Does Pe sonali y In luence…
75
Table 4.3. Va imax o a ed componen ma ix o he p oposed ou is a ac ions, showing he 11 ac o s ex ac ed using EFA,
he es ima ed co ela ions be ween he i ems and ac o s, and each ac o s C onbach’s Alpha eliabili y.
Fac o
I em
Desc ip ion
Es ima ed co ela ions
α
Ad en u e
F1
A46
P ac ice climbing o bungee jumping
0,775
0.876
A63
Do ai spo s (e.g., pa achu e jump, skydi ing, gliding)
0,758
A68
Ski
0,750
A12
P ac ice aqua ic spo s (e.g., sailing, canoeing, di ing, je
skiing)
0,744
A29
Do mo o spo s (e.g., ka ing, mo oc oss)
0,627
A37
Obse e sub-aqua ic en i onmen s / ma ine li e (e.g., sno kel-
ing, subma ine)
0,575
Na u e
F2
A41
Walk in he o es / woods
0,821
0.857
A11
Do hiking / moun ainee ing
0,721
A42
Take a walk along he i e / seacoas
0,662
A47
Visi moun ain a eas / go ges
0,635
A36
Visi na u e o wildli e ese es
0,632
A10
App ecia e na u al landscapes
0,625
A6
Visi ca es/ca e ns/ olcanoes
0,523
En e ainmen &
Nigh li e
F3
A44
Go o a dance/balle es i al
0,823
0.849
A43
Go o a music es i al/conce
0,743
A45
Go o balls (dancing)
0,722
A48
Go o a li e music ba /place
0,683
A17
Go o a ilm es i al
0,624
A67
Assis o an ope a/ hea e
0,621
A8
A end cul u al ac i i ies / a is ic pe o mances
0,512
A9
Go o he disco/nigh club
0,473
Sun, Wa e & Sand
F4
A65
Go o he swimming pool o elax
0,797
0.807
A22
Go o he beach (sunba hing/ swimming)
0,725
A66
Ha e aca ion on an island
0,648
A64
Go o he swimming pool o swim/di e
0,642
Museums & Land-
scapes
F5
A32
Visi museums o his o ical hemes
0,741
0.775
A34
Visi iewpoin s o na u al landscape
0,731
A33
Visi museums o scien i ic hemes (e.g., plane a ium, paleon-
ology)
0,670
A35
Visi iewpoin s o u ban landscape
0,637
Themes & Animal
Pa ks
F6
A15
Go o a Zoo
0,821
0.736
A13
Go o a heme pa k (e.g., Disneyland Pa is)
0,641
A27
Go o a wa e pa k
0,590
Cul u al He i age
F7
A7
Visi a chaeological si es / uins
0,737
0.763
A20
Visi monumen s (e.g. chu ches, ca hed als, cas les, o ess-
es, monas e ies, palaces, e c.)
0,720
A4
Visi he his o ic ci ies/ illages o he des ina ion
0,657
Spo s & Games
F8
A56
Assis o a spo ing compe i ion (e.g., wa ch a oo ball game
om a club o ha coun y)
0,764
0.723
A55
Play a he casino
0,638
A53
Play ball spo s (e.g., oo ball, handball, olleyball, ennis)
0,601
A59
Pa icipa e in an escape game
0,572
Gas onomy
F9
A1
Go o a Gas onomy Fes i al ( ood and/o d inks)
0,844
0.742
A40
Pa icipa e in a gas onomy ou ( ypical and/o gou me
dishes, wine as ing)
0,772
A18
Tas e ypical local dishes
0,674
Boa Tou s
A51
Take boa ips o he pleasu e o boa ing
0,626
0.790
Chap e 4 (Pape 2) - Modeling Tou is s' Pe sonali y in Recommende Sys ems: How Does Pe sonali y In luence…
76
Fac o
I em
Desc ip ion
Es ima ed co ela ions
α
F10
A49
Take boa ips o know he des ina ion’s coas
0,619
A50
Take boa ips o he his o ical alue o he ou e
0,555
Heal h & Well-being
F11
A28
Go o a SPA / beau y cen e
0,582
0.733
A14
Unde go heal h and wellness ea men s (e.g., hyd o he apy
cen e s, mine al wa e eso s)
0,575
A26
A end gyms / i ness cen e s
0,556
Ac i i ies like going o he beach o a swimming pool a e s ongly ela ed o mo e ex a e ed, open
o expe ience and neu o ic indi iduals, as also ound by Delic, Neidha d , and We hne (2016).
Ex a e s may app ecia e he oppo uni y o be in con ac wi h o he people, bo h new and known
acquain ances, and a he same ime eel com o able wi h hei body-image. Indi iduals wi h high
openness may enjoy he aes he ic expe ience ha comes wi h app ecia ing a beach o i s na u al
beau y, along wi h he medi a i e s a e his su oundings may induce, con i ming he esul s ound
by Bujisic e al. (2015). In u n, neu o ics may app ecia e hese ac i i ies o consis ing in some-
wha p edic able aca ions, almos a de aul op ion in he con ex o ou coun y, hus eassu ing a
s ong sense o sa e y and con ol which is dea o hese indi iduals.
Table 4.4. Pe sonali y dimensions mean o al sco es (BFI-44), n=508.
Mean
SD
Min/Min ange
Max/Max ange
Ex a e sion
24
6
10/8
40/40
Ag eeableness
33
5
14/9
45/45
Conscien iousness
32
5
11/9
45/45
Neu o icism
25
6
10/8
39/40
Openness
37
6
18/10
50/50
Visi ing museums and landscape iewing a e p e e ed by mo e open o expe ience indi iduals,
while ex a e ed a e nega i ely associa ed. This goes in line wi h he indings o Bujisic e al.
(2015) and Jani (2014b). Sense can also be made in ligh o he ac ha enjoying his o ical a i-
ac s o pu suing in ima e expe iences commonly b ough by abso bing impac ul scene ies a e
ac i i ies mo e akin o indi idual enjoymen and soli ude.
The p e e ence o isi ing Theme and Animal Pa ks is posi i ely p edic ed by pe sons wi h highe
Ag eeableness (p obably ela ed o he ones who ge along wi h he ones who enjoy hose ac i i-
ies, such as child en and adolescen s), Neu o icism ( he same as he ag eeable in e p e a ion)
and Ex a e sion ( ela ed o ene ge ic and exci emen seeking pe sons). The p e e ence o Theme
and Animal Pa ks is nega i ely p edic ed by Openness o expe ience, possibly o i s s anda dized,
no -so-in ellec ually-challenging na u e. No signi ican p edic ion was ound o Conscien iousness.
Chap e 4 (Pape 2) - Modeling Tou is s' Pe sonali y in Recommende Sys ems: How Does Pe sonali y In luence…
77
A cheological si es/ uins, monumen s and his o ic ci ies/ illages a e also posi i ely sough by
ag eeable, conscien ious and open o expe ience indi iduals, bu nega i ely by ex a e s, who may
no ind e y exci ing ha so o ac i i ies as hey a e a e se o seden a ism, con i ming he esul s
ound by Jani (2014b), bu con adic o y o he esul s ound by Delic, Neidha d , and We hne
(2016). This con adic ion can be due o he di e en cul u es s udied o he sample size, al hough
o he easons can be he cause.
Spo s and games a e nega i ely associa ed o conscien ious and open o expe ience indi iduals,
bu ag eeableness and ex a e sion a e posi i ely ela ed, which can de i e om he coope a i e-
ness, ene gy, and high ac i i y le el inhe en o hese ac i i ies. This con i ms wha was ound o
Ad en u e p e e ences, and he indings o Schneide and Vog (2012), who ound compe i i eness
was associa ed o so ad en u e a ele s.
Gas onomy expe iences a e posi i ely alued by mo e ag eeable and ex a e ed people, which
can explain why who app ecia es ood and wine is usually seen as chee ul and high-spi i ed. No
signi ican ela ions we e ound o he o he h ee pe sonali y ai s.
Boa ou s a e p e e ed by indi iduals wi h highe ag eeableness and openness o expe ience,
being in line wi h he esul s ound by Jani (2014b), which is simila o he in e p e a ion gi en o
beach/swimming pool ac i i ies and landscape iewing.
Heal h and well-being a e alued by hose wi h highe Ag eeableness and Ex a e sion, bu wi h
lowe Openness. No signi ican associa ion was ound o Conscien iousness and Neu o icism,
which does no mean hey do no exis .
The esul s ound demons a e a s ong ela ion be ween he i e pe sonali y dimensions and p e -
e ences o ou is a ac ions, i.e., p e e ences o speci ic ou is a ac ions can be p edic ed by
he ou is s’ pe sonali y meaning hey can be used in RS o ou ism o au oma ically model he
use s p e e ences.
4.5
R
EFLECTIONS AND
F
UTURE
W
ORK
In he domain o a el and ou ism, li le in o ma ion is known abou he p edic ion o ou is p e -
e ences based on he ou is s’ pe sonali y. This wo k b ough new insigh s, success ully and
s ongly ela ing all he Big Fi e pe sonali y dimensions o p e e ences o speci ic ou is a ac-
Chap e 4 (Pape 2) - Modeling Tou is s' Pe sonali y in Recommende Sys ems: How Does Pe sonali y In luence…
78
ions, showing all he ex ac ed ou is a ac ions ac o s a e ele an and could be p edic ed by
he pe sonali y dimensions, cons i u ing, o he bes o ou knowledge, he i s “Pe sonali y-Tou is
A ac ions P e e ence” model in li e a u e o do so.
Al hough he model was con i med wi h a sa is ac o y i , and since his s udy is pa o an ongoing
la ge s udy, we belie e he model’s i will conside ably imp o e wi h he inc ease in size and he -
e ogenei y o he sample, p o ided he used sample was small o he numbe o a iables o es i-
ma e.
I was also ound ha due o he sel - epo ing answe s in he BFI, some pe sonali y dimensions
had a highe concen a ion abo e mid-poin sco es, which may e lec some social desi abili y bias,
o ha he sample should be mo e di e se. In ac , mos o he esponden s had a highe educa-
ion le el, which can explain he g ea e equency o highe sco es in dimensions like Openness o
Expe ience and Conscien iousness, meaning his s udy is limi ed in p edic ing ou is a ac ions
p e e ences o indi iduals wi h lowe educa ion. Also, mos esponden s we e om Exac o Social
sciences, meaning he sample was poo on o he ields o educa ion, which can also accoun o
he lowe a iabili y ound in Openness o Expe ience, Conscien iousness and Ag eeableness,
since, o ins ance, indi iduals wi h A is ic o ma ion would supposedly posi i ely in luence he
Openness o Expe ience o e all sco e.
Also, we canno o ge he ques ionnai e was conduc ed in a Po uguese popula ion, and he e o e
some ela ionships ound be ween pe sonali y and ou is a ac ions can be cul u e speci ic. How-
e e , ou in en ion is o conduc he same s udy on di e en cul u es and compa e he esul s.
Al hough pe sonali y e ealed o be a g ea p edic o o ou is p e e ences, o ocus only on he
i e pe sonali y dimensions and no on he mo e g anula hi y ai s and on he co ela ions be-
ween hem, may limi he p e e ences p edic ion. Fo example, a pe son conside ed ex a e ed
may no be a isk ake o like ad enaline- illed ac i i ies. I would no be e y good i he RS sug-
ges ed a olle coas e o he ou is . Since he ques ionnai e allowed o ob ain signi ican da a, i
will be u he analyzed, along wi h he incoming da a, o de e mine which o he 30 pe sonali y
ai s can be in e ed om he BFI.

Chap e 4 (Pape 2) - Modeling Tou is s' Pe sonali y in Recommende Sys ems: How Does Pe sonali y In luence…
79
I is also e idenced ha he pe sonali y s uc u e a ies subs an ially ac oss cul u es (McC ae &
Cos a J , 1997), and ha ce ain pe sonali y ai s a e mo e lean o show di e en emo ions ac-
co ding o he social con ex hey a e in (Odić, Tkalčič, Tasič, & Koši , 2013). The e o e, o know
he pe sonali y is no enough, all hese a iables need o be co ela ed o p o ide highe quali y
ecommenda ions and a g ea e use expe ience.
O he psychological aspec s, such as mood, in en ions, emo ions, and mo i a ions ha e also
shown o in luence he ou is s’ p e e ences (Odić e al., 2013; Tkalčič e al., 2016). The e o e,
o he psychological aspec s will also be a ocus in ou u u e wo k.
I is impo an o e e ha , in addi ion o psychological ac o s, o he ac o s, like he geog aphic
and cul u al dis ance should be included in he s udies o he ela ionship be ween he ou is s’
pe sonali y and des ina ion choice (McKe che , 2006; K. Y. Poon & Huang, 2017).
I is qui e e iden ha pe sonali y is no he only esponsible o ou is p e e ences, bu i seems
undeniable ha pe sonali y plays an impo an ole in he kind o ou is a ac ions one chooses o
isi .
Acknowledgmen s
This wo k was suppo ed by he G ouPlanne P ojec unde he Eu opean Regional De elopmen
Fund POCI-01-0145-FEDER-29178 and by Na ional Funds h ough he FCT – Fundação pa a a
Ciência e a Tecnologia (Po uguese Founda ion o Science and Technology) wi hin he P ojec s
UIDB/00319/2020 and UIDB/00760/2020.
80
5
G R O U P R E C O M M E N D E R S Y S T E M S F O R T O U R I S M :
H O W D O E S P E R S O N A L I T Y P R E D I C T P R E F E -
R E N C E S F O R A T T R A C T I O N S , T R A V E L M O T I V A -
T I O N S , P R E F E R E N C E S A N D C O N C E R N S ?
An image showing di e en des ina ion possibili ies and ou is a ac ions.
Gene a ed wi h AI, Image C ea o in Bing, Oc obe 2024
81
Pape Ti le
G oup ecommende sys ems o ou ism: how does pe sonal-
i y p edic p e e ences o a ac ions, a el mo i a ions,
p e e ences and conce ns?
Au ho s
Pa ícia Al es, Helena Ma ins, Ped o Sa ai a, João Ca nei o, Paulo
No ais, Go e i Ma ei os
Pape Type
Jou nal A icle
Jou nal Name
Use Modeling and Use -Adap ed In e ac ion (UMUAI)
Yea
2023
DOI
h ps://doi.o g/10.1007/s11257-023-09361-2
Published
15 May 2023
Publishe
Sp inge
URL
h ps://link.sp inge .com/a icle/10.1007/s11257-023-09361-2
JCR 2023
IF 3.0 in: Compu e Science, Cybe ne ics #10/32 Q2
Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
82
Abs ac
To a el in leisu e is an emo ional expe ience, and he e o e, he mo e in o ma ion abou he ou -
is is known, he mo e pe sonalized ecommenda ions o places and a ac ions can be made. Bu
i o p o ide ecommenda ions o a ou is is complex, o p o ide hem o a g oup is e en mo e.
The eme gence o pe sonali y compu ing and pe sonali y-awa e ecommende sys ems (RS)
b ough a new solu ion o he cold-s a p oblem inhe en o he con en ional RS, and can be he
le e age needed o sol e con lic ing p e e ences in he e ogenous g oups and o make mo e p ecise
and pe sonalized ecommenda ions o ou is s, as i has been e idenced ha pe sonali y is s ong-
ly ela ed o p e e ences in many domains, including ou ism. Al hough many s udies on psycholo-
gy o ou ism can be ound, no many p edic he ou is s’ p e e ences based on he Big Fi e pe -
sonali y dimensions. This wo k aims o ind how pe sonali y ela es o he choice o a wide ange o
ou is a ac ions, a elling mo i a ions, and a el- ela ed p e e ences and conce ns, hoping o
p o ide a solid base o esea che s in he ou ism RS a ea o au oma ically model ou is s in he
sys em wi hou he need o edious con igu a ions, and sol e he cold-s a p oblem and con lic ing
p e e ences. By pe o ming Explo a o y and Con i ma o y Fac o Analysis on he da a ga he ed
om an online ques ionnai e, sen o Po uguese indi iduals om di e en a eas o o ma ion and
age g oups (
𝑛
=1035), we show all i e pe sonali y dimensions can help p edic he choice o ou -
is a ac ions and a el- ela ed p e e ences and conce ns, and ha only neu o icism and open-
ness p edic a elling mo i a ions.
Keywo ds:
G oup Recommende Sys ems; Pe sonali y; Tou is P e e ences; T a el Mo i a ions;
T a el Conce ns; A ec i e Compu ing
5.1
I
NTRODUCTION
The las wo decades ha e shown ha pe sonaliza ion is he key o deli e he bes ecommenda-
ions in Recommende Sys ems (RS) (L. Chen e al., 2016; K. Y. Poon & Huang, 2017; Tkalcic &
Chen, 2015; Tondello e al., 2017). The e o e, he mo e abou he use is known he mo e accu-
a e and ailo ed ecommenda ions can be made. Bu i o p o ide ailo ed indi idual ecommenda-
ions is complex, o p o ide hem o g oups is e en mo e (Mas ho , 2015). The ou ism indus y
has many a iables making i a e y complex opic, which is agg a a ed when g oups o ou is s,
Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
89
M. Jackson e al. (2001) p oposed ou ypes o ou is s: he explo e , he ad en u e , he guided,
and he g oupie, combining he o hogonal scales o Allocen ics-Psychocen ics and In o e sion-
Ex a e sion.
Eachus (2004) p oposed a Holiday P e e ences Scale by modi ying bo h Plog’s and M. Jackson e
al. (2001) ypologies: Ad en u ous p e e ence, Beach p e e ence, Cul u al p e e ence, and Indul-
gen p e e ence. They ound ha ou is s wi h high sco es in sensa ion seeking ended o p e e
Ad en u ous and Beach holidays bu no Indulgen holidays, and olde ou is s we e mo e likely o
p e e Cul u al holidays han younge .
Based on Cohen’s indi idual mass ou is ype (E. Cohen, 1972), Wickens (2002) p oposed i e
mic o- ypes o ou is s: he Cul u al He i age, he Ra e , he Shi ley Valen ine, and he Lo d By on
ype.
To enhance he ele ance o ecommenda ions in RS, G e zel e al. (2004) p oposed 12 a el
pe sonali ies and s udied how hey ela ed o 17 a el ac i i ies in No he n Indiana, ha ing ound
s ong co ela ions be ween hem. The mos selec ed a el pe sonali ies we e All A ounde , Sigh
Seeke and Cul u e C ea u e. Conce ning he ela ionships ound, as also la e e i ied by Delic,
Neidha d , and We hne (2016), mos a el pe sonali ies we e ela ed o mul iple ac i i ies, o
ins ance, Shopping Sha ks ype was ela ed o ou is s mo e in e es ed in shopping, nigh li e, and
dining. Cul u e C ea u es p e e ed es i als, museums, and his o ic si es. Family Guy was no
ela ed o gambling, biking, o hun ing/ ishing. T ail T ekke s we e less ela ed o Ci y Slicke ,
Shopping Sha k, and Game a el ypes. Boa e s did no conside hemsel es as Sigh Seeke s,
and Beach Bums did no iden i y wi h he His o y Bu ca ego y. The o he ypes co esponded o
hei espec i e ac i i ies. La e , he same au ho s s udied i he p oposed a el pe sonali ies could
p edic he ac i i ies and/o places o be ecommended by a des ina ion RS (G e zel e al., 2006),
inding a el pe sonali ies a e “ e y good p oxies o cap u ing use pe sonali y ai s and p e e -
ences and can be used o make speci ic des ina ion ecommenda ions” (G e zel e al., 2006).
As poin ed by G e zel e al. (2006), “i is no clea how easy i is o indi iduals o selec and iden i-
y wi h an exis ing” ypology o how hey can ac ually p edic he ou is s’ beha io . Al hough being
a po en ial way o desc ibing ypes o ou is s and c ea ing ma ke ing segmen s, ypologies do no
allow o unde s and wha pe sonali y dimensions and/o ai s a e behind hose p e e ences, and

Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
90
he e o e a e no easy o implemen in a (G)RS wi hou he need o ce ain ini ial con igu a ions by
he use , p oblem we p opose o mi iga e by au oma ically p edic ing he ou is s’ p e e ences o
ou is a ac ions based on hei pe sonali y.
Pe sonali y as P edic o o P e e ences o Tou is A ac ions/Des ina ions
As poin ed by se e al au ho s, he exis ing esea ch on ou ism beha io is mos ly desc ip i e in-
s ead o p edic i e (M. Jackson e al., 2001; Schneide & Vog , 2012) which is a limi a ion ha
needs o be o e come, i.e., wha pe sonali y dimensions o ai s a e p edic i e o he ou is s’
ypologies o beha io s/p e e ences ound in li e a u e?
C o s (1990) ound he mo e dogma ic (close-minded) he pa icipan s we e, he less no el y and
mo e amilia i y hey wan ed in hei aca ions, and he ones ha had a g ea e need o cogni ion,
and endency o engage in and enjoy hinking sough mo e o no el y.
Acco ding o Bujisic e al. (2015), indi iduals wi h highe le el o openness o expe ience ended o
be mo e sa is ied wi h aes he ic and escapis expe iences han hose wi h lowe le el. In con as ,
indi iduals wi h lowe openness o expe ience we e mo e sa is ied wi h en e ainmen and educa-
ional expe iences compa ed o he ones wi h highe le el. Ex o e s ended o be mo e sa is ied
wi h educa ional and escapis expe iences.
Al hough many s udies on psychology o ou ism o di e en ou ism sec o s can be ound, many
a e abou ypologies o ou is s (Addison, 1999; Lipscombe, 1995; Milling on e al., 2001; S. C.
Plog, 2002), which as men ioned be o e, a e desc ip i e o he ou is s’ beha io and do no p e-
dic how ha beha io in luences he choice o ou is p e e ences. O he s y o p edic how psy-
chological aspec s in luence ou is beha io o p e e ences, bu mos o hem only ely on Sensa-
ion Seeking, ex a e sion, and/o openness o expe ience scales (Bujisic e al., 2015; M. Jackson
e al., 2001; C.-Y. Li e al., 2015; Nicke son & Ellis, 1991), which do no co e all Big Fi e’s di-
mensions. Few s udies y o co ela e all Big Fi e dimensions wi h ou is beha io s o p e e -
ences. Fo example, Neidha d , Schus e , Sey ang, and We hne (2014); Neidha d , Sey ang,
Schus e , and We hne (2015) pe o med a ac o analysis on he 17 ou is oles p oposed by
Gibson and Yiannakis (2002) and he Big Fi e pe sonali y dimensions, ob aining se en ac o s ha
cap u ed he ou is s beha io , whe e some o hem e ealed o be co ela ed wi h pe sonali y di-
Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
91
mensions: (1)
Sun lo ing and connec ed
– highly co ela ed o he sun lo e ou is ole and he
neu o icism, openness and conscien iousness pe sonali y dimensions; (2)
Educa ional
– co ela es
o ganized mass ou is and educa ional ou is wi h ag eeableness; (3)
Independen
– combines
independen mass ou is s I and II and seeke ; (4)
Cul u e lo ing
– co ela es a chaeologis and
high-class ou is wi h ex a e sion; (5)
Open minded and spo i e
– combines an h opologis and
spo ou is wi h ex a e sion; (6)
Risk seeking
– esul s om he co ela ion o ac ion seeke , ex-
plo e and je se e ; (7)
Na u e and silence lo ing
– co ela es escapis I and II.
K aso a (2015) s udied how pe sonali y in luenced ou is s’ eco- iendly beha io , inding indi idu-
als wi h high ag eeableness we e s ongly ela ed o eco- iendly beha io , ollowed by conscien-
iousness and neu o icism, con i ming se e al pas s udies on he same a ea o esea ch (Hi sh,
2010; Ma kowi z e al., 2012; Mil on & Sibley, 2012). Rega ding openness o expe ience, indi id-
uals wi h high imagina ion we e nega i ely associa ed o eco- iendliness bu indi iduals wi h high
in ellec we e posi i ely associa ed.
Jani (2014b) and Delic, Neidha d , and We hne (2016) s udied how he Big Fi e co ela ed o a
a ie y o ou is oles. Jani (2014b) explo ed ha ela ionship using he Big Fi e and he 12 a el
pe sonali ies ( ypes) p oposed by G e zel e al. (2004). The au ho ound signi ican pe sonali y
di e ences be ween he a el ypes. Those who enjoy games o any ype (A hle e ype), his o ical
si es, shopping, and wa e ac i i ies/a ac ions (Boa e ) a e high in openness o expe ience, and
hose who like o lay a ound he beach (Beach bum) and spend ime wi h amily a e low in ha
dimension. Shopping and Family a el ypes ha e a high conscien iousness, and A hle e and
Game ypes a e low in ha ac o . T ekke and All hings a el ypes ha e highe ex a e sion,
and Cul u al, Beach bum, and Boa e ypes a e lowe in ex a e sion. As o high ag eeableness, i
includes Boa e and Family a el ypes, and low ag eeableness he Game ype. Low neu o icism
was associa ed wi h Family and All hings a el ypes. Delic, Neidha d , and We hne (2016) ana-
lyzed he ela ionship be ween he 17 ou is oles de ined by Gibson and Yiannakis (2002) and he
Big Fi e. Fo example, Sun Lo e ype was ela ed o high neu o icism indi iduals, A cheologis o
ex a e s, and D i e o less conscien ious people. No signi ican co ela ions we e ound be ween
he o he ou is oles. As expec ed, hey also ound ou is oles a ied wi h age, bu ha he Big
Fi e pe sonali y dimensions we e s able ac oss ages.
Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
92
All hese s udies show ha he a el beha io and p e e ences a e ela ed o he ou is s’ pe sonal-
i y. Howe e , none, o he bes o ou knowledge, co ela es he Big Fi e pe sonali y ac o s o he
choice o aw ca ego ies o ou is a ac ions, bu ins ead, o ou is ypologies. Wi h his wo k, we
in end o ill ha gap by p oposing a model o p edic he p e e ence o a wide ange o ou is
a ac ions, adap ed om he “Thesau us on Tou ism and Leisu e Ac i i ies” o he Wo ld Tou ism
O ganiza ion (UNWTO, 2001), based on he ou is s’ i e pe sonali y dimensions, aiming o help
ou ism (G)RS o p o ide ecommenda ions o isi ing a ac ions/des ina ions jus by knowing he
ou is ’s pe sonali y, and sol e p oblems ela ed o con lic ing p e e ences in (occasional excu sion)
g oups. We belie e ha c ea ing subg oups wi h simila pe sonali ies, and he e o e, simila ou is
p e e ences, can help sol e hose con lic s.
This esea ch is mo i a ed by he e idence ound in li e a u e, om which i is possible o eason
ha he ou is ypologies do no ully jus i y he ou is s’ p e e ences o ou is a ac ions, since
many di e en combina ions o in ensi y o he pe sonali y ai s exis and he e o e a single ypol-
ogy may no be enough o a ce ain ou is as well as no all he a ac ions p esen in a ypology
may be sui able o ha ou is . This claim is suppo ed by he esul s ound by G e zel e al.
(2004) and Neidha d e al. (2015). Al hough i is “easy” o ecommend a ac ions based on a
single pe sonali y dimension, indi iduals ha e a combina ion and di e en sco es on he i e pe -
sonali y dimensions. How do we agg ega e all ha o ecommend he igh ou is a ac ions? We
canno ecommend an a ac ion classi ied o high ex a e sion and low neu o icism o someone
low in bo h dimensions.
5.2.2.2
Tou ism Mo i a ion
Many s udies on ou ism mo i a ion exis , some s udying mo i es o a elling o speci ic si es
(Collins-K eine & Klio , 2000), ou ism niches (Hassani & Mogha emi, 2019; Heung & Leong,
2006; O oo & Kim, 2020), senio ou is s (Boksbe ge & Laesse , 2008; O oo, Kim, Ag usa, &
Lema, 2021; Pa uelli & Nijkamp, 2016; Vigolo, 2017), o in gene al (Hei mann, 2011), o he s o
p opose scales o dimensions o mo i a ions (C omp on, 1979; Pea ce & Cal abiano, 1983),
among o he s. These s udies a e pa icula ly impo an o ou ism ma ke ing, and he e o e o ou -
ism (G)RS, so be e and mo e ailo ed se ices and p oduc s can be deli e ed o ou is s.
Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
93
One o he i s esea che s o ca e o he ou is s’ mo i a ions was Dann (1977), by ying o an-
swe he ques ion “Wha makes ou is s a el?”. Al hough some iewpoin s could be ound a he
ime, claiming he majo eason o a elling was o escape om he daily ou ine, he o dina y,
e c., he e was no empi ical e idence o demons a e i (Dann, 1977; Lundbe g, 1971). Howe e ,
a gene al classi ica ion o explain ou is mo i a ion wi h “push” and “pull” ac o s was widely ac-
cep ed (Dann, 1977; Hei mann, 2011). “Push” ac o s e e o he ou is ’s physiological and psy-
chological aspec s (e.g., escape, elax, e c.) in luencing his decision o a el, like needs and p e -
e ences. “Pull” ac o s pe ain o he cha ac e is ics o he a el des ina ion o ex e nal mo i a-
ions ha a ac (pull) he ou is o isi i .
La e , Iso-Ahola (1982), sugges ed ou ism mo i a ion was mainly d i en by escape and seeking,
bo h ha ing pe sonal (psychological) and in e pe sonal (social) ac o s, and he e o e he dis in-
guished ou dimensions: pe sonal seeking, pe sonal escape, in e pe sonal seeking, and in e pe -
sonal escape.
McIn osh and Goeldne (1985) p oposed i e ypes o mo i a ions, e lec ing he ideas o he
Maslow’s py amid: Physical ( he need o elaxa ion o o he ac i i ies o educe s ess o e esh
he body and mind), Emo ional ( o seek omance, ad en u e, spi i uali y, escapism o nos algia),
Cul u al ( o lea n abou he des ina ion’s cul u e and he i age), In e pe sonal ( he need o main ain
o de elop new ela ionships, by isi ing ela i es o iends, o mee new people), and S a us and
p es ige ( he need o enhance sel -s a us and ecei e a en ion/ alo iza ion om o he s).
A e y in e es ing a el mo i a ion heo y was de eloped by Pea ce (1993), Pea ce and Cal abiano
(1983), and Mosca do and Pea ce (1986): he T a el Ca ee Ladde (TCL), la e modi ied o T a -
el Ca ee Pa e n (TCP) since ou is s could be a mo e han one le el a a ime (Pea ce & Lee,
2005). Also based on he Maslow’s needs hie a chy (Maslow, 1970), he heo y desc ibes i e
di e en le els o ou is needs, om bo om o op: elaxa ion needs, sa e y/secu i y needs, ela-
ionship needs, sel -es eem and de elopmen needs, and inally, sel -ac ualiza ion/ ul illmen
needs. The heo y a gues ha ou is s’ mo i a ion changes acco ding o hei age and/o a el
expe ience, esul ing in a a el ca ee . To unde s and pleasu e a el mo i a ion mo e b oadly,
Pea ce and Lee (2005) iden i ied a wide ange o a el mo i e i ems and de e mined 14 unde ly-
ing mo i a ion ac o s: No el y, Escape/ elax, Rela ionship (s eng hen), Au onomy, Na u e seek-
Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
94
ing, Sel -de elopmen (hos -si e in ol emen ), S imula ion, Sel -de elopmen (pe sonal de elop-
men ), Rela ionship (secu i y), Sel -ac ualize, Isola ion, Nos algia, Romance, and Recogni ion. They
ound escape/ elax, no el y, ela ionship, and sel -de elopmen we e he mos impo an mo i es
o a elling. The mo e expe ienced a ele s we e mo e mo i a ed by sel -de elopmen h ough
hos -si e in ol emen and na u e seeking. The low expe ienced we e mo e d i en by s imula ion,
pe sonal de elopmen , sel -ac ualiza ion, secu i y, nos algia, omance, and ecogni ion.
Li e a u e on a el mo i a ion is e y ex ensi e and he e o e only some wo ks we e p esen ed, bu
one hing is clea , he main easons o a elling ha e been e y simila in he las decades and
among di e en age echelons, being Explo a ion, o ha e Cul u al/Na u e expe iences, and Relaxa-
ion/Escapism he mos common mo i es.
Pe sonali y as P edic o o Tou ism Mo i a ion
By analyzing why someone chooses o a el o a speci ic si e o ou is a ac ion can help ind
hei a elling and beha io al pa e ns, which can g ea ly help imp o e he ou is ’s p o ile in a
(G)RS and hus p o ide be e ecommenda ions. And, i , o ins ance, pe sonali y could be ela ed
o he mo i es behind a elling, i would be easie o p opose ce ain a ac ions o des ina ions by
jus knowing he ou is ’s pe sonali y. As Hei mann (2011) poin s ou , many ac o s can in luence
he ou is s’ beha io and choices, such as cul u al and eligious ac o s, demog aphics, and pe -
sonal ac o s such as pe sonali y, li es yle, occupa ion, income, among o he s. So, how does pe -
sonali y ela e o he mos common ou is s’ mo i a ions?
As men ioned in Sec ion 5.2.2, se e al ou is oles and ypologies ha e been p oposed o desc ibe
ou is beha io s (E. Cohen, 1972; G ay, 1970; S. C. Plog, 1974; V. L. Smi h, 2012) bu hey do
no explain he easons behind hose beha io s (Hei mann, 2011).
No many wo ks ha s udy he ela ionship be ween (Big Fi e) pe sonali y and a elling mo i a-
ions we e ound, and he ones exis ing, o he bes o ou knowledge, aim o ela e pe sonali y and
he mo i a ions o speci ic ou ism niches o des ina ions, such as eligious ou ism and c uise
ship ou is s (Abba e & Di Nuo o, 2013; Sca idi Abba e, Di Nuo o, & Ga o, 2017), a el cu iosi y
(Jani, 2014a; Kashdan e al., 2009), olun ee ou ism (Suhud, 2015), o jus o he a el desi e
(Labbe, 2016), o ela e o he pe sonali y ypes o he a el in en ion (Kaewumpai, 2017; H. Kim,

Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
95
Yilmaz, & Choe, 2019; Kwon & Pa k, 2015; O oo e al., 2021), o only o one mo i e. O he s use
di e en scales o pe sonali y (no he Big Fi e).
A ecen s udy o Sca idi Abba e e al. (2017) compa ed he mo i a ions (
Cu iosi y and disco e y
,
Ou -o - ou ine
and
Sel and sociali y
) and pe sonali y o eligious a ele s e sus c uise ship ou -
is s. Rega ding eligious a ele s, he au ho s ound openness o expe ience posi i ely p edic ed
Cu iosi y and disco e y
mo i a ion and ag eeableness nega i ely. Ag eeableness and conscien-
iousness nega i ely p edic ed
Ou -o - ou ine
mo i a ion.
Sel and sociali y
was p edic ed by nega-
i e sco es in openness o expe ience. A di e en pa e n was ound in c uise ou is s, whe e
openness o expe ience and ag eeableness bo h posi i ely in luenced he cu iosi y mo i a ion, and
conscien iousness nega i ely.
Ou -o - ou ine
mo i a ion was nega i ely p edic ed by conscien ious-
ness and neu o icism. Finally, openness o expe ience, ex a e sion (ene gy) and conscien ious-
ness posi i ely p edic ed
Sel and sociali y
mo i a ion.
The indings in li e a u e show ha ce ain mo i a ions o a elling in speci ic con ex s can be
p edic ed by pe sonali y dimensions. Wi h his wo k, we in end o e i y i ha applies o a g ea e
ange o a el mo i a ions, including he mos common ones, abs ac ed om speci ic niches o
des ina ions, and o p opose a model o p edic ou ism mo i a ions based on he ou is s’ Big Fi e
pe sonali y dimensions.
5.2.2.3
T a el-Rela ed P e e ences and Conce ns
To choose a a el des ina ion is pa o a p ocess ha s a s wi h he need/desi e o a elling
(Ma hieson & Wall, 1982), and he in o ma ion ha is collec ed is e alua ed acco ding o he a -
ele ’s needs and p e e ences as well as possible cons ain s. Acco ding o Hung e al. (2016),
he e a e h ee ypes o a el cons ain s: in ape sonal (e.g., o eel guil y o a elling, o be
a aid o a elling o a speci ic des ina ion, limi ed knowledge o ou ism), in e pe sonal (e.g., lack
o a el pa ne s), and s uc u al (e.g., lack o ime o money). Fo ins ance, many people would
like o isi Uk aine, bu due o he ac ual wa i is no a choice. Also, someone migh p e e o isi
a coun y on summe ins ead o ano he season. O someone may no be able o a el due o
money o ime cons ain s. In his s udy, we ocused in he in ape sonal and in e pe sonal con-
s ain s and will conside hem as conce ns om now on.
Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
96
Some conce ns ha e shown o in ensi y wi h age (Fleische & Pizam, 2002; Lindq is & Bjö k,
2000; Vigolo, 2017; You & O'Lea y, 1999), like he ea o becoming ill, lack o doc o a ailabili y,
conce ns o sa e y and pe sonal secu i y, sani a ion, se ice and ood quali y (J. Kim, Wei, & Ruys,
2003; Lindq is & Bjö k, 2000; To es & Skillico n, 2004). Heal h p oblems a e mo e no iceable in
olde ou is s (> 64 yea s old), educing he leng h o aca ions (Fleische & Pizam, 2002), and
inc easing he conce ns abou a elling o long-haul o less de eloped des ina ions, ligh du a-
ions, heal h insu ance, o e en humidi y (Hun e
‐
Jones & Blackbu n, 2007). As poin ed by Vigolo
(2017), Huang and Tsai (2003) ound senio Taiwanese a ele s e ealed p eoccupa ion o lea -
ing hei house una ended, no ha ing a el companions, die a y es ic ions, o no ha ing an
enjoyable ime and was e money. Chinese women we e mo e conce ned abou “limi ed knowledge
o ou ism, heal h and sa e y, cul u e shock, lack o a el pa ne s, low quali y se ice acili ies,
limi ed a ailabili y o in o ma ion, and nega i e epu a ion o ou guide” (Gao & Ke s e e , 2016;
Vigolo, 2017). Emo ional ba ie s like ea o he unknown, loss o eedom and loss o spon anei y
we e poin ed as he highes ba ie s o amily ca egi e s and hei ca e- ecipien s by Gladwell and
Bedini (2004).
Al hough sa e y and secu i y ha e long been key conce ns o many ou is s (La sen, B un, &
Øgaa d, 2009; A. Poon & Adams, 2000), ou ism in gene al is no seen as isky (Sönmez &
G ae e, 1998a, 1998b). Howe e , ce ain unexpec ed and agical e en s can dec ease he ou -
is s’ con idence and educe he desi e o a el. The a acks on he Wo ld T ade Cen e and he
Pen agon, on Sep embe 11, 2001, we e a sad example, which led o he mass cancela ion o
inbound and ou bound ligh s (Floyd, Gibson, Penning on-G ay, & Thapa, 2004). The ac ual COVID-
19 pandemic is ano he case, whe e o a el, ei he o e seas o wi hin he same coun y, is con-
side ed isky, and was e en o bidden o many coun ies (Bo kowski, Ja
ż
d
ż
ewska-Gu a, &
Szmel e -Ja osz, 2021; Godo ykh, Pizam, & Bahja, 2021; Mo a e al., 2021; Neubu ge & Egge ,
2021; Tabak, Canik, & Gune en, 2021; Zenke , B aun, & Gyimó hy, 2021).
The s udy o he pe cei ed isks in ou ism has long been in es iga ed (Dolnica , 2005), being he
concep i s in oduced by Baue (1960). Acco ding o Dolnica (2005), he s udy o pe cei ed
isks can be classi ied in o wo dimensions: nega i e pe cei ed isks, which a e no sough by he
ou is , and posi i e pe cei ed isks, which a e ac i ely sough by he ou is , such as sensa ion
Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
97
seeking ac i i ies. In hei in es iga ion o he ea s Aus alian ou is s associa e o leisu e a el, in
he con ex o domes ic and o e seas a el, Dolnica (2005) ound i e ca ego ies o isk ac o s:
(1) poli ical isk, such as “ e o ism, poli ical ins abili y, wa /mili a y con lic ”; (2) en i onmen al
isk, like “na u al disas e s, landslides”; (3) heal h isk, like “lack o access o heal h ca e, li e
h ea ening diseases, lack o access o clean ood and wa e ”; (4) planning isk, such as “un elia-
ble ai line, inexpe ienced ope a o , no assu ed ligh home”; and (5) p ope y isk, such as “ he ,
loss o luggage”. All he e e ed isks we e mo e equen ly associa ed o o e seas a el. As o
domes ic a el, wildli e and he oad’s condi ion we e he g ea es conce ns. The con ex may
change e e y hing. Nega i e e en s in associa ion o ea s and conce ns can p e en he ou is
om isi ing ce ain places/a ac ions, om being in ol ed in pa icula ac i i ies, o e en om
a elling.
As o a el- ela ed p e e ences, someone migh p e e o a el accompanied o alone, o a cold
o ho wea he des ina ion, o o ake a ligh o a el by ca , and so on. T a el p e e ences can
also in luence which des ina ion o isi o e en he decision o a el a all. Fo ins ance, O oo,
Kim, and Pa k (2020) s udied eigh a el- ela ed ea u es/p e e ences: a el du a ion (by ligh ),
a el pa ne s, accommoda ion ype, a el a angemen ype (own o package ou ), in o ma ion
echnology accep ance, ou ism ype (e.g., u ban, eco, heal h), a ac ions ype (e.g., his o ical,
na u al scene y), and ac i i ies ype (ou doo , shopping, dining), and ela ed hem o he a el
mo i a ions hey ound. Rami es, B andao, and Sousa (2018) s udied wha a el p e e ences and
des ina ion a ibu es ou is s isi ing Po o in Po ugal p e e ed, namely: a el o ganize , a el
pa ne s, anspo o des ina ion, ype o accommoda ion, ype o ac i i ies in he des ina ion,
anspo in he des ina ion, and how hey we e ela ed o hei a el mo i a ions.
Many a el p e e ences and conce ns can in luence he a el plans, b inging limi a ions o e en
p e en ou is s om a elling. To know hem can imp o e (G)RS ecommenda ions. Bu how does
pe sonali y i s in hese all? Is i an in luencing ac o o hose p e e ences and/o conce ns?
Pe sonali y as P edic o o T a el-Rela ed P e e ences and Conce ns
Many s udies ha ela e ou is ypologies o pe sonali y o a el- ela ed p e e ences and/o con-
ce ns, especially conce ns, could be ound.
Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
98
Sand a Lee Basala (1997); Sand a L Basala and Klenosky (2001) ound indi iduals wi h di e en
a el s yles (Familia i y Seeke s, A e age T a ele s, and No el y Seeke s) had di e en a el-
ela ed p e e ences, namely ega ding he ype o accommoda ions, ype o a el companions,
and language o he hos des ina ion.
Beside ela ing demog aphics, M. Jackson and R. Inbaka an (2006) s udied ou is s isi ing Aus-
alia and how hei p oposed pe sonali y ypes (Explo e : in o e +allocen ic, Ad en u e : ex a-
e +allocen ic, Guided: psychocen ic+in o e , and G oupie: psychocen ic+ex a e ) ela ed o
he p e e ence o p e-planning a aca ion, using in e ne o book a els, a elling alone, a el
companions, in en ion o e isi a des ina ion, leng h o s ay, and des ina ion’s cul u al simila i y.
Conside ing h ee pe sonali y dimensions (ex a e sion, conscien iousness, and emo ional sensi i -
i y (neu o icism)), Ma i z, Yeh, and Shieh (2013) s udied how ou is s’ pe sonali y in luenced he
pe cei ed a el isks (pe sonal isk, p ope y isk and liabili y isk), he a elling in en ion, and he
pe cei ed isk on a el in en ion. As o pe cei ed a el isk, conscien iousness and emo ional
sensi i i y posi i ely in luenced pe sonal isk, emo ional sensi i i y posi i ely p edic ed p ope y isk,
and inally, all h ee pe sonali y dimensions showed posi i e e ec s on liabili y isk. Ex a e sion
did no appea o be a ec ed by a el isks. Rega ding a el in en ions, all h ee pe sonali y di-
mensions posi i ely in luenced he a el in en ion. Pe cei ed isk signi ican ly a ec ed a el in en-
ion in e ms o pe sonal and liabili y isks. E idence sugges ed he pe cei ed isk would educe he
in luence o ex a e sion, conscien iousness, and emo ional sensi i i y on a el in en ion.
Using Plog’s psychog aphic model (S. C. Plog, 2002), Mo akaba i and Kapu
ś
ci
ń
ski (2016) ocused
on he ela ionship be ween isk pe cep ion and he des ina ions’ bene i s, and i e o ism a ec ed
he willingness o a el acco ding o he ou is s’ pe sonali y. Yazdanpanah and Hosseinlou
(2016), s udied he in luence o he Big Fi e pe sonali y dimensions on he choice o anspo a ion
mode acco ding o he wea he condi ions.
Ca alho, Pianowski, and Gonçal es (2020) s udied i ex a e sion and conscien iousness we e
ela ed o social dis ancing and handwashing COVID-19 con ainmen measu es in B azil. They
e i ied he less ex a e ed he mo e conce ned wi h social dis ancing he pa icipan s we e. Pa -
icipan s who conside ed nei he o he wo con ainmen measu es had lowe conscien iousness
Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
105
n
%
n
%
Eu ope, Ame ica
142
13.7
Las ime
isi ed
ano he
coun y
Las 6 mon hs
425
41.1
Eu ope, Asia,
A ica, Ame ica
87
8.4
Las 2 yea s
389
37.6
O he combina ions
109
11,0
Mo e han 2 yea s
ago
188
18.2
Ne e isi ed o he
con inen s
29
2.8
Ne e
33
3,1
T a el
companions
F iends/Colleagues
158
15.3
T a els
ab oad pe
yea
Ne e
171
16.5
Pa ne
231
22.3
3 imes o less
802
77.5
Pa ne and
child en
277
26.8
4 o 6 imes
54
5.2
Rela i es
312
30.1
7 o 10 imes
4
0.4
Nobody
41
4.0
Mo e han 10 imes
4
0.4
O he
16
1.5
Some s a is ics a e no shown as hey a e ela ed o ano he ongoing s udy.
Compa ed o he p e ious s udy (Al es e al., 2020), he e a e mo e adul s be ween 23 and 55
yea s old (56% be o e, now 70%), and he numbe o esponden s wi h child en inc eased om
31% o 42%, meaning he e a e mo e pa icipan s supposedly wi h di e en esponsibil-
i y/conce ns. The o ma ion a eas a e also mo e a ied. The o he sample cha ac e is ics emain
simila .
(a) (b)
(c) (d)
376
55
219
67
25
284
9
Be ween 1001 € -2000 €
Be ween 2001 € -3000 €
Be ween 650 € -1000 €
Less han 650 €
Mo e han 3000 €
No applicable
Re use o answe
214
324
57
130
108
55
51
271
14
Accoun ing and Finances
Ag icul u al Sciences
Enginee ing and Technology
Exac Sciences
Humani ies
Medical and Heal h Sciences
Na u al Sciences
None
Social Sciences
Unknown
802
54
4
4
171
3 imes o less
4 o 6 imes
7 o 10 imes
Mo e han 10 imes
Ne e
Figu e 5.1. Responden s’ (a) age ange, (b) o ma ion a eas, (c) liquid income, (d) leisu e a els ab oad pe yea ;
𝑛=1035

Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
106
5.4.1.2
Pe sonali y
To assess he esponden s’ pe sonali y, he Big Fi e In en o y (44 i ems) was used, which is one
o he mos widely used pe sonali y in en o ies. The BFI assesses an indi idual on he Goldbe g’s
Big Fi e dimension o pe sonali y (Lewis R Goldbe g, 1990), using a 5-poin Like scale. To acili-
a e in e p e a ion, ins ead o calcula ing he sco es o each pa icipan , he mean alue o each
pe sonali y dimension is p esen ed. Figu e 5.2 shows he esponses dis ibu ion o each dimen-
sion.
Figu e 5.2. Dis ibu ion o he i e pe sonali y dimensions esponses (pa icipan s’ mean alue)
Clea ly, he e a e 3 dimensions wi h esponses abo e he mid-poin , e ealing a sligh nega i e
skewness, i.e., pa icipan s si ua ed hemsel es mo e be ween “3-Nei he ag ee no disag ee” and
Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
107
“5-Ag ee s ongly”: ag eeableness, conscien iousness and openness, con i ming he esul s ound
in ou p e ious s udy (Al es e al., 2020), e lec ing he same social desi abili y bias, which usually
happens mo e in sel - epo ing ques ionnai es (Ped egon e al., 2012), like he desi e o being kind
and mo al in he case o ag eeableness; u h ul, sel -e ec i e and e o ul in he case o conscien-
iousness; and mo e in ellec ual in he case o openness o expe ience. The o he wo dimensions,
ex a e sion and neu o icism, ha e he mean alue nea he scale mid-poin .
All i e dis ibu ions ollow he shape o a no mal cu e, and acco ding o he alues o skewness
and ku osis ob ained, and espec i e s anda d e o s, al hough ha ing a sligh skew and ku osis,
hey a e in accep able anges and he da a is conside ed no signi ican ly di e en om a no mal
dis ibu ion
23
(Field, 2013; G a e e , Wallnau, Fo zano, & Wi naue , 2020; Sposi o, Hand, &
Ska pness, 1983).
5.4.1.3
Tou is a ac ions p e e ence
A o al o 68 i ems ep esen ing a wide ange o di e en ou is a ac ions, ollowing he mos
signi ican e ms om he Uni ed Na ions Wo ld Tou ism UNWTO (2001), we e p esen ed o he
esponden s in he ques ionnai e. Table 5.3 summa izes he agg ega ed esul s.
Table 5.3. Pa icipan s' p e e ences o ou is a ac ions, in pe cen age o ag eemen .
To ally
Disag ee
To ally
Ag ee
1
2
3
4
5
6
7
A1
Go o a Gas onomy Fes i al ( ood and/o d inks)
5.7
6.6
7.0
14.1
24.5
22.2
19.9
A2
Wa ch a na u al phenomenon (e.g., olcanic
e up ion o no he n ligh s)
2.6
2.0
3.5
7.6
12.7
25.9
45.7
A3
Wa ch a eligious celeb a ion
21.1
13.6
13.6
18.3
15.9
9.9
7.6
A4
Visi he his o ic ci ies/ illages o he des ina ion
0.5
1.3
1.7
5.6
13.1
30.6
47.1
A5
Visi an oceana ium
7.6
5.8
8.2
17.9
23.4
20.6
16.5
A6
Visi ca es/ca e ns/ olcanoes
4.3
3.1
3.4
10.8
19.1
27.5
31.8
A7
Visi a chaeological si es / uins
1.4
2.4
3.1
12.0
20.3
26.3
34.6
A8
A end cul u al ac i i ies / a is ic pe o mances
1.4
1.7
4.3
11.1
23.3
30.6
27.5
A9
Go o he disco/nigh club
28.3
16.0
12.9
16.0
13.6
8.3
4.7
A10
App ecia e na u al landscapes
0.2
0.3
0.8
3.0
8.7
27.7
59.3
A11
Do hiking / moun ainee ing
2.7
3.1
5.8
10.0
24.9
24.4
29.0
A12
P ac ice aqua ic spo s (e.g., sailing, canoeing,
di ing, je skiing)
12.6
10.1
10.6
17.8
16.0
15.5
17.4
A13
Go o a heme pa k (e.g., Disneyland Pa is)
6.4
6.8
6.4
10.7
19.3
22.7
27.7
A14
Unde go heal h and wellness ea men s (e.g.,
hyd o he apy cen e s, mine al wa e eso s)
10.6
8.1
9.6
18.1
21.1
15.0
17.6
23
The s a is ics o he BFI esponses can be ound in Appendix B a h p://www.gecad.isep.ipp.p /g ouplanne /dissemina ion.h ml (UMUAI Appendix).
Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
108
To ally
Disag ee
To ally
Ag ee
1
2
3
4
5
6
7
A15
Go o a Zoo
15.6
11.1
7.9
18.1
18.7
14.9
13.7
A16
A end a ypical celeb a ion o he des ina ion (e.g.,
popula celeb a ions, ca ni als, i ewo ks)
2.4
1.5
3.3
8.9
19.8
31.5
32.6
A17
Go o a ilm es i al
10.3
9.4
12.2
21.0
21.0
14.8
11.4
A18
Tas e ypical local dishes
0.7
1.0
1.6
5.6
14.0
25.7
51.4
A19
Visi a bo anical ga den
4.5
4.3
5.2
16.4
24.4
23.9
21.3
A20
Visi monumen s (e.g. chu ches, ca hed als, cas les,
o esses, monas e ies, palaces, e c.)
1.0
1.5
3.4
5.7
15.1
26.9
46.5
A21
Visi a beach o i s na u al beau y
1.4
0.3
0.8
4.2
11.0
27.0
55.5
A22
Go o he beach (sunba hing/ swimming)
5.4
3.6
4.6
7.9
17.7
21.3
39.5
A23
To enjoy / buy local handic a s
2.4
2.5
4.2
14.7
27.0
23.8
25.5
A24
Ride a bike
11.3
7.5
8.9
19.8
22.4
16.5
13.5
A25
Go o a un ai (e.g., amusemen s such as Fe is
wheel, bumpe ca s, e c.)
14.0
10.0
9.9
18.6
19.0
14.8
13.7
A26
A end gyms / i ness cen e s
44.2
17.6
13.6
11.1
7.1
3.9
2.6
A27
Go o a wa e pa k
16.9
10.0
11.2
15.9
16.2
13.5
16.2
A28
Go o a SPA / beau y cen e
22.1
11.6
11.6
15.7
14.7
12.0
12.4
A29
Do mo o spo s (e.g., ka ing, mo oc oss)
30.5
14.0
11.4
14.8
12.2
8.9
8.2
A30
Ha e a picnic
6.7
5.6
6.3
17.3
24.9
21.3
18.0
A31
To go shopping / see s o e on s (window shopping)
18.4
13.6
11.5
18.6
18.7
10.5
8.7
A32
Visi museums o his o ical hemes
3.8
7.3
10.5
9.5
19.1
22.5
27.2
A33
Visi museums o scien i ic hemes (e.g.,
plane a ium, paleon ology)
3.8
6.3
10.8
11.2
19.3
24.0
24.6
A34
Visi iewpoin s o na u al landscape
0.5
2.4
7.0
6.9
11.5
24.7
47.1
A35
Visi iewpoin s o u ban landscape
3.7
7.5
10.3
14.7
20.4
21.3
22.1
A36
Visi na u e o wildli e ese es
1.8
1.3
2.8
7.4
15.5
28.8
42.4
A37
Obse e sub-aqua ic en i onmen s / ma ine li e
(e.g., sno keling, subma ine)
10.3
6.2
8.4
14.1
15.3
20.2
25.5
A38
Visi la ge man-made cons uc ions (e.g., b idges,
unnels, mines)
6.3
6.1
11.3
16.4
22.1
19.5
18.3
A39
Go o a hema ic pa ade (e.g., mili a y, elec onic
music)
19.0
13.8
16.1
20.3
16.3
8.4
6.0
A40
Pa icipa e in a gas onomy ou ( ypical and/o
gou me dishes, wine as ing)
5.9
6.7
8.5
14.0
24.4
20.9
19.6
A41
Walk in he o es / woods
3.3
3.1
3.2
11.2
24.3
28.1
26.8
A42
Take a walk along he i e / sea coas
1.1
1.2
2.2
6.3
20.4
32.1
36.8
A43
Go o a music es i al/conce
8.5
6.6
9.8
15.3
21.2
21.5
17.2
A44
Go o a dance/balle es i al
15.0
11.4
12.9
18.4
17.6
14.0
10.8
A45
Go o balls (dancing)
24.6
14.3
15.7
16.7
12.8
8.2
7.7
A46
P ac ice climbing o bungee jumping
30.2
13.3
13.2
13.8
13.7
6.5
9.2
A47
Visi moun ain a eas / go ges
7.4
7.1
10.0
14.1
21.7
20.0
19.5
A48
Go o a li e music ba /place
7.7
5.6
9.2
17.4
23.6
21.2
15.4
A49
Take boa ips o know he des ina ion's coas
4.1
5.0
7.4
11.5
20.4
25.0
26.6
A50
Take boa ips o he his o ical alue o he ou e
5.2
6.5
8.2
13.3
20.7
24.0
22.1
A51
Take boa ips o he pleasu e o boa ing
12.9
10.0
10.8
14.7
15.7
18.2
17.8
A52
Take a walk in a ci y pa k
1.0
2.0
2.5
11.4
25.1
33.1
24.8
A53
Play ball spo s (e.g., oo ball, handball, olleyball,
ennis)
33.6
16.4
11.6
13.5
11.3
6.4
7.1
A54
Do a sa a i
10.0
6.4
7.0
11.5
20.4
19.2
25.6
A55
Play a he casino
52.2
15.1
10.8
9.7
6.4
3.6
2.3
Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
109
To ally
Disag ee
To ally
Ag ee
1
2
3
4
5
6
7
A56
Assis o a spo ing compe i ion (e.g., wa ch a
oo ball game om a club o ha coun y)
34.8
12.0
10.3
13.5
13.2
8.7
7.4
A57
Ride a ho se
25.3
11.9
10.7
15.3
15.7
10.8
10.2
A58
Hun / ish
61.4
13.2
7.5
7.8
4.3
2.9
2.8
A59
Pa icipa e in an escape game
40.4
13.6
9.9
13.7
9.6
6.0
6.9
A60
Wa ch a bull igh
82.0
6.3
3.6
4.0
2.1
0.8
1.3
A61
Go o he ci cus
53.8
12.6
9.0
10.4
7.1
3.7
3.5
A62
Go on a c uise
14.4
7.3
6.7
14.6
17.2
17.6
22.2
A63
Do ai spo s (e.g., pa achu e jump, skydi ing,
gliding)
33.9
11.2
9.2
11.8
11.6
10.0
12.4
A64
Go o he swimming pool o swim/di e
11.6
5.4
7.3
13.2
18.6
19.4
24.4
A65
Go o he swimming pool o elax
5.3
4.9
5.6
10.5
18.8
23.1
31.7
A66
Ha e aca ion on an island
3.2
2.3
2.9
8.1
18.5
27.2
37.8
A67
Assis an ope a/ hea e
13.3
7.5
8.4
15.3
20.1
19.2
16.1
A68
Ski
28.9
9.5
10.2
16.0
13.6
8.8
12.9
The highes alues ha con ibu e o mos o he esponses a e i alics, and he alues ha a e simila ly dis ibu ed a e unde lined. Values >45% a e in bold.
E e yone seems o like almos all so o ou is a ac ions when a elling on aca ions, wi h a
clea signi ican majo i y on a ac ions like wa ching a na u al phenomenon, isi his o ic ci -
ies/ illages, app ecia e na u al landscapes (including beau i ul beaches), as e ypical local dishes,
and isi monumen s. The opposi e can also be said o a ending gyms, playing a he casino,
hun ing/ ishing, (decidedly) wa ching bull igh s, and going o he ci cus, which a e de ini ely no a
choice when on aca ion. Wha ela ionship exis be ween hose p e e ences and he pa icipan s’
pe sonali y? Sec ion 5.4.2.2 shows he esul s.
5.4.1.4
T a elling mo i a ions
The pa icipan s’ a elling mo i a ions we e measu ed using Pea ce and Lee (2005) p oposed
i ems, using he wo i ems wi h highes loading o each mo i e. The i ems we e hen mixed up in
he ques ionnai e’s espec i e sec ion so i ems om he same mo i e ended sepa a ed. The ag-
g ega ed esul s can be ound in Table 5.4.
Clea ly, excep o ques ions M8 and M22, pa icipan s a e on he same side ega ding mo i a ions
o a elling in leisu e, con i ming mos mo i es p oposed by Pea ce and Lee (2005). To be close
o na u e, mee he locals, ha e ad en u esome expe iences, de elop pe sonal in e es s and skills,
unde s and mo e abou sel & wo k on pe sonal alues, be wi h espec ul pe sons, ge isola ed,
eel he des ina ion’s a mosphe e, expe ience some hing di e en , ge away om e e yday
s ess/demands, in e ac wi h amily/ iends & s eng hen hose ela ionships, ha e no obliga ions
Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
110
& be independen , a e all possible mo i es o a elling in leisu e. The g ea majo i y also ag eed
ha o mee new amo ous pa ne s and ge ecognized by o he s we e no mo i es o go on aca-
ion. I is ob ious ha no all mo i es a e sui able o he same ype o aca ions and need o be
con ex ualized. And is he e a simila pe sonali y be ween simila a elling mo i es? In Sec ion
5.4.2.3 we analyze how pe sonali y ela es o hose mo i es.
As p e iously men ioned, he pa icipan s esponded qui e di e en ly o wo ques ions measu ing
Pea ce and Lee (2005) Nos algia dimension, M8 and M22. Mos o hem ag eed hey wan o hink
abou good imes spen in he pas bu a e equally di ided in e lec ing on pas memo ies (37% o
ag ee and disag ee). Al hough sco ing in he same Nos algia dimension, acco ding o he esul s
ob ained, M8 is ela ed o hinking abou good memo ies, and M22 ela es o pas memo ies, ei-
he good o bad, possibly sugges ing e lec ions akin o lea ning om expe ience (Table 5.4).
5.4.1.5
T a el- ela ed p e e ences and conce ns
One sec ion o he ques ionnai e was ela ed o a el- ela ed p e e ences and conce ns (Al es e
al., 2020), whe e we asked he pa icipan s ques ions ela ed o hei p e e ences and conce ns
when a elling. The ques ions and agg ega ed esul s a e shown in Table 5.5.
Much in o ma ion can be ob ained om he esponses collec ed, bu only he ele an o his
s udy is p esen ed. Mos esponden s: p e e ou doo ac i i ies bu a e no willing o ake physical
isks; like o s udy he des ina ion’s his o y p io o a elling bu conside i is no impo an o plan
he aca ion days in ad ance and ha he e should be no ime schedules; wan he des ina ion o
include cul u al/lea ning componen s and y o include as many a ac ions as possible; a e no
wo ied i he e is no mobile phone ne wo k a ailable; like des ina ions whe e ew people ha e
been o, conside ing impo an o see exo ic a ac ions o di e en om hei cul u e, bu would
ne e isi an impo an ci y wi hou seeing i s iconic monumen s, no eeling mo e keen o isi a
des ina ion o being “in ogue” o media ized; would no a el o a highly pollu ed o high c imi-
nali y/a med con lic s des ina ion; conside impo an he accommoda ion’s com o ; always buy
sou eni s; would accep a a el package om a a el agency bu would like o be in ol ed in he
choice p ocess; a e no incommoded i hey ha e o spend aca ions wi h s ange s o a el in a
g oup o ganized by a a el agency, bu would p e e o a el wi h ou is s simila o hem, all co -
esponding o a da a dis ibu ion wi h posi i e o nega i e skewness. I is impo an o no ice ha

Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
111
al hough hese esul s ep esen he median ou is , hey do no mean ha o each ou is he
p e e ences clus e his way.
Table 5.4. Pa icipan s' a elling mo i a ions, in pe cen age o ag eemen (ques ions adap ed om Pea ce and Lee (2005)).
To ally
Disag ee
To ally
Ag ee
1
2
3
4
5
6
7
M1
Being close o na u e
1.1
1.1
3.1
9.3
19.5
29.5
36.5
M2
Mee ing he locals
1.7
2.4
4.8
12.9
23.7
28.8
25.6
M3
Ha ing da ing/ad en u esome expe ience
3.4
3.5
7.6
16.9
24.3
24.7
19.6
M4
De elop my pe sonal in e es s
0.1
0.3
0.8
5.9
17.9
36.3
38.7
M5
Being wi h espec ul people
0.4
0.3
0.9
8.4
12.4
29.7
48.0
M6
Unde s anding mo e abou mysel
1.4
1.8
2.6
16.3
18.1
26.2
33.6
M7
Being away om he c owds o people
3.3
5.9
9.2
26.8
20.7
16.5
17.7
M8
Thinking abou good imes I’ e had in he pas
4.5
6.6
7.4
23.3
18.2
19.1
20.9
M9
Ha ing oman ic ela ionships
25.4
13.2
7.0
20.6
11.8
11.5
10.5
M10
Showing o he s I can do i
22.3
13.4
11.2
23.8
11.5
9.6
8.2
M11
Feeling he special a mosphe e o he aca ion
des ina ion
0.5
0.6
1.0
6.6
17.4
32.6
41.4
M12
Ge ing away om e e yday psychological
s ess/p essu e
0.5
0.7
1.5
4.3
9.8
25.5
57.8
M13
Doing some hing wi h my amily/ iend(s)
0.3
0.1
0.3
4.8
11.0
27.7
55.7
M14
Being obliga ed o no one
3.5
3.4
4.0
13.3
13.9
20.6
41.4
M15
Ge ing a be e app ecia ion o na u e
0.7
1.6
1.4
7.6
16.9
28.1
43.6
M16
Obse ing o he people in he a ea
3.2
3.8
4.7
17.1
21.2
23.5
26.6
M17
Expe iencing h ills
3.9
3.9
7.7
18.6
25.6
18.3
22.0
M18
De eloping my skills and abili ies
1.7
1.3
4.0
14.5
21.6
26.4
30.5
M19
Being nea conside a e people
1.3
1.7
2.4
13.4
20.8
28.1
32.3
M20
Wo king on my pe sonal/spi i ual alues
2.8
2.7
5.0
20.6
19.7
24.0
25.2
M21
Enjoying isola ion
7.4
10.0
10.5
23.6
20.8
14.4
13.3
M22
Re lec ing on pas memo ies
11.6
12.4
13.0
26.0
14.6
10.8
11.6
M23
Mee ing amo ous pa ne s
53.8
15.5
5.7
13.7
4.8
3.8
2.7
M24
Being ecognized by o he people
23.9
11.7
11.6
21.7
14.4
8.7
8.0
M25
Expe iencing some hing di e en
0.9
1.2
1.3
8.6
20.9
28.8
38.5
M26
Ge ing away om he usual demands o li e
0.9
1.1
1.6
6.5
11.2
26.1
52.7
M27
S eng hening ela ionships wi h my
amily/ iend(s)
1.4
0.4
1.2
6.7
14.1
28.5
47.7
M28
Being independen
3.8
2.8
3.8
17.7
17.4
22.8
31.8
The highes alues ha con ibu e o mos o he esponses a e i alics, and he alues ha a e simila ly dis ibu ed a e unde lined. Values >45% a e in bold
Cu iously, he e is a clea “balanced” di ision in some esponses: 40% o he esponden s would
p e e o dine wi h a mos 5 people and 43% would no ; 41% is no a aid o ge ing ill o acciden s
while on aca ions and 46% is; 33% ca e abou he money spen on aca ion and 31% do no ca e;
41% ag ee i is always be e o a el in g oup, ega dless o he des ina ion, and 40% do no ; 41%
a e no a aid o ge ing los in a dis an coun y, bu 45% a e; 32% p e e o ha e a good su p ise
while 28% p e e o ul ill expec a ions; 30% a e indi e en i hey con ibu e o he des ina ion’s
local economy bu 34% like o eel hey con ibu e. Rega ding he des ina ion’s wea he condi ions,
Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
112
35% p e e ho wea he , bu also 35% do no ha e a p e e ence, and 28% like wa m wea he . Cold
wea he is no a choice o he esponden s. Mos pa icipan s ha e some so o phobias/ ea s
(e.g., ea o heigh s, con ined spaces, e c.).
P obably no all p e e ences and conce ns we e chosen by he same ype o pa icipan s. Do pe -
sonali y dimensions p edic a el- ela ed p e e ences and conce ns? I so, which ones? In Sec ion
5.4.2.4, we answe hose ques ions.
Table 5.5. Pa icipan s' a el- ela ed p e e ences and conce ns, in pe cen age o ag eemen .
To ally
Disag ee
To ally
Ag ee
1
2
3
4
5
6
7
P1
When a elling on leisu e, I
p e e ou doo ac i i ies
1.1
2.1
4.0
11.0
22.4
30.5
28.9
P2
Unde no ci cums ances I like
o ake isks ela ed o my
physical in eg i y
3.8
9.3
13.6
14.4
15.8
18.4
24.7
P3
A dinne wi h iends ideally
should ha e a maximum o 6
people
16.3
13.8
12.4
17.5
15.6
13.1
11.3
P4
When going on aca ions I
ake in o accoun he
des ina ion's cul u al o e
1.5
3.0
5.4
12.5
16.2
29.8
31.6
P5
I am a aid o ge ing ill o
ha ing acciden s whils away
on aca ions
11.8
17.6
11.4
13.6
17.2
14.8
13.6
P6
I would ne e a el o a place
whe e he e was no mobile
phone ne wo k
21.4
18.5
14.7
14.9
10.4
10.5
9.6
P7
When planning aca ions, I y
o include as many
places/a ac ions as possible
2.0
4.9
7.1
13.8
18.4
24.3
29.5
P8
To be pe ec , a aca ion
needs ha e e y day is
planned in ad ance
14.5
16.3
17.5
15.7
16.3
12.1
7.5
P9
Wi hin my possibili ies, when
on a aca ion I don' look a
expenses
10.8
14.8
18.6
16.4
17.1
13.5
8.8
P10
Be o e a elling I like o
know/s udy he his o y o he
des ina ion
3.1
5.2
9.6
14.1
22.9
24.2
21.0
P11
Rega dless o he des ina ion,
i is always be e o a el in
g oup
10.9
14.3
14.8
19.5
15.6
12.7
12.3
P12
In a dis an coun y, one o
my wo s ea s would be o
ge los
12.8
14.6
13.7
13.7
15.0
15.6
14.7
P13
I like o go whe e ew people
ha e been o be o e
7.7
9.6
11.6
23.6
16.7
17.1
13.7
Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
113
To ally
Disag ee
To ally
Ag ee
1
2
3
4
5
6
7
P14
Fo me, o eel com o is
always he mos impo an
(quali y o acili ies/p oduc s)
3.3
7.0
11.7
17.3
29.4
19.4
12.0
P15
Fo me, o ul ill expec a ions
is mo e impo an han a
good su p ise
8.7
14.1
17.6
25.7
15.2
12.7
6.1
P16
When on aca ions ab oad, I
like o eel ha I am
con ibu ing o he local
economy
8.2
8.2
10.1
30.3
19.2
15.1
8.8
P17
Fo me, while on aca ions
he e should be no ime
schedules
3.2
7.1
11.9
14.9
18.4
20.3
24.3
P18
Fo me, a good aca ion has
o include a cul u al/lea ning
componen
1.9
3.9
5.1
14.0
24.3
25.8
25.0
P19
I would ne e isi a g ea ci y
wi hou seeing i s iconic
monumen s
3.9
4.3
7.0
10.2
18.9
23.8
31.9
P20
I would ne e a el o a
des ina ion wi h high pollu ion
le els
4.6
10.3
18.3
19.8
15.9
17.4
13.6
P21
Fo me, i 's impo an o see
exo ic hings o ha a e e y
di e en om my cul u e
1.5
4.9
6.7
16.5
20.8
26.0
23.6
P22
When I e u n om a
aca ion, I always b ing
sou eni s o me, amily o
iends
4.2
6.8
4.4
10.0
16.5
22.3
35.7
P23
I a des ina ion is "in ogue"
o appea s in he media, I eel
mo e like isi ing i
14.6
14.7
17.0
22.1
16.4
9.4
5.8
P24
I would be willing o a el in
a g oup o ganized by a a el
agency
8.4
10.2
11.5
15.8
17.6
18.2
18.3
P25
I would ne e go on aca ion
wi h s ange s (making
common ips and meals)
16.3
17.4
17.9
16.5
11.7
10.7
9.5
P26
I I we e o a el in a g oup, I
would a he do i wi h people
simila o me
3.0
3.9
7.3
20.1
23.9
23.6
18.3
P27
I would ne e a el in a
g oup due o p i acy easons
34.2
21.4
15.4
14.1
6.6
4.0
4.4
P28
I would be incapable o
a elling o a high c iminali y
a e / a med con lic
des ina ion
2.9
6.9
7.6
9.7
13.7
21.5
37.7
Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
114
To ally
Disag ee
To ally
Ag ee
1
2
3
4
5
6
7
P29
I like o isi uncommon
places o obse e peculia
hings (e.g., wo ld eco ds,
pop icons, his o ical i ems,
e c.)
2.2
4.6
8.2
17.9
24.9
23.7
18.5
P30
I'm a aid o :
P31
When planning aca ions, I gene ally
p e e o go a place wi h:
Heigh s
30.9
Cold
wea he
2.5
T a elling on wa e
3.4
Wa m
wea he
27.6
Flying
3.6
Ho
wea he
35.0
Being unde wa e
11.6
I don'
ha e a
p e e ence
34.9
Con ined spaces
14.1
O he
5.4
I ha e no ea s
31.0
P34
Conside ing an i ine a y /
aca ion plan p esen ed by a
a el agency, I would p e e :
To ha e a comple e
p oposal, wi h e e y hing
de ined, ' eady o use'
24.3
To be in ol ed in he
choice p ocess, ha e
mo e con ol and moni o
all s ages o he p ocess
75.7
The highes alues ha con ibu e o mos o he esponses a e i alics, and he alues ha a e simila ly dis ibu ed a e unde lined
Some ques ions a e no shown as hey a e ela ed o ano he ongoing s udy.
5.4.2
How does Pe sonali y p edic P e e ences o Tou is A ac ions, T a el Mo i a ions, P e -
e ences and Conce ns?
In his sec ion, we p esen he esul s o he EFA and CFA pe o med on he ques ionnai e i ems
o each s udied a el aspec , excep o pe sonali y, whe e we p esen only he CFA esul s.
5.4.2.1
Pe sonali y
The EFA o he BFI esponses con i med he Big Fi e pe sonali y dimensions, agg ega ing he
i ems in o he expec ed pe sonali y dimensions. CFA con i med he EFA esul s (Table 5.6), bu
some i ems we e emo ed due o a eg ession weigh < 0.45 (we conside ed i ems abo e his
h eshold as he scale consis ency inc eased), esul ing in 21 i ems om he 44 in he o iginal
scale, as he i ems used in he p oposed models had o ep esen he sample used o he s udy.
This p obably means he sample needed o be la ge o main ain all he 44 i ems.
All he dimensions’ C onbach’s Alpha c ossed he 0.60 h eshold o psychological a iables (John
& Bene -Ma ínez, 2000), ha ing an accep able o good eliabili y (Geo ge & Malle y, 2019). The
Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
121
wi h amily and lay a he beach. Neu o icism can be ela ed o he need o p edic able aca ions,
which can be ound in ypical beach/ho el- ela ed aca ions.
Table 5.8. S anda dized eg ession weigh s o he ela ionship be ween he BFI dimensions and he p e e ence o Tou is
A ac ions, ob ained using CFA.
Tou ism Ca ego y
BFI Dimension
Reg ession Weigh
p
Ad enaline Ac i i ies
F1
Ex a e sion
0.715
***
Conscien iousness
-0.320
***
Ag eeableness
0.024
0.524
Neu o icism
0.012
0.706
Openness
-0.039
0.224
Wild Na u e Ac i i ies
F2
Ex a e sion
0.404
***
Ag eeableness
0.573
***
Conscien iousness
-0.223
***
Neu o icism
0.053
0.120
Openness
0.017
0.608
Pa y, Music & Nigh li e
F3
Ex a e sion
0.751
***
Ag eeableness
-0.050
***
Neu o icism
0.129
***
Openness
-0.115
***
Conscien iousness
-0.108
0.003**
Sun, Wa e & Sand
F4
Ex a e sion
0.617
***
Neu o icism
0.076
0.026*
Openness
-0.232
***
Ag eeableness
0.008
0.827
Conscien iousness
0.016
0.648
Museums, Boa ips & Viewpoin s
F5
Ex a e sion
0.063
0.097
Ag eeableness
0.525
***
Neu o icism
0.078
0.033*
Openness
0.078
0.029*
Conscien iousness
-0.182
***
Theme & Animal Pa ks
F6
Ex a e sion
0.790
***
Ag eeableness
-0.123
0.003**
Neu o icism
0.128
***
Openness
-0.204
***
Conscien iousness
-0.077
0.026*
Cul u al He i age
F7
Ag eeableness
0.625
***
Openness
-0.006
0.085
Ex a e sion
-0.019
0.612
Neu o icism
0.044
0.213
Conscien iousness
-0.006
0.864
Spo s & Games
F8
Ex a e sion
0.717
***
Ag eeableness
-0.309
***
Openness
-0.152
***
Conscien iousness
-0.150
***
Neu o icism
0.010
0.796
Gas onomy E en s
F9
Ex a e sion
0.459
***
Ag eeableness
0.187
***
Openness
-0.116
0.002**
Conscien iousness
-0.089
0.023*
Neu o icism
-0.010
0.784

Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
122
Tou ism Ca ego y
BFI Dimension
Reg ession Weigh
p
Heal h & Well-being
F10
Ex a e sion
0.649
***
Ag eeableness
-0.168
***
Neu o icism
0.144
***
Openness
-0.143
***
Conscien iousness
0.079
0.029*
Na u al Phenomena
F11
Ex a e sion
0.336
***
Ag eeableness
0.605
***
Conscien iousness
-0.365
***
Neu o icism
0.045
0.246
Openness
0.020
0.598
S a is ically signi ican alues a e in bold and shaded in g een (* p < 0.05 (2- ailed), ** p < 0.01 (2- ailed), *** p < 0.001 (2- ailed))
The ca ego ies Museums & Landscapes and Boa Tou s om he p e ious s udy me ged in o he
new ca ego y Museums, Boa ips & Viewpoin s, d opping he i em ela ed o boa ing jus o he
pleasu e o i . The EFA in he inc eased sample allowed o ind simila i ies be ween he wo ac o s,
p oposing hey should belong o he same ac o . They all ha e in common o iew/app ecia e
some na u al o his o ical scene y, which was no he case o boa ing o pleasu e. These p e e -
ences we e ound o be p edic ed by ou pe sonali y dimensions, wi h a s onge in luence om a
posi i e ag eeableness and nega i e conscien iousness, p obably because less cau ious people a e
mo e willing o ake boa ips, and since iewpoin s a e gene ally loca ed in high places, high con-
scien ious pe sons may no be willing o go as hey migh be mo e suscep ible o heigh s. The o h-
e wo dimensions sligh ly posi i ely in luenced he p e e ence, ha ing a small ele ance in he
p edic ion.
Going o a wa e pa k, un ai s, zoo o heme pa ks a e ac i i ies p e e ed by highly ex a e ed
pe sons (ene ge ic and exci emen -seeking), sligh ly neu o ic (indi iduals ha eel mo e com o a-
ble wi h iends and amily), wi h nega i e ag eeableness, openness (indi iduals who p e e amili-
a i y, s anda d and no so in ellec ually challenging ac i i ies), and conscien iousness ( e ealing
pe sons ha a e willing o ake mino isks). Wi h he new da a, we obse ed a wis in he impac
o ag eeableness, which now is nega i e.
Monumen s and his o ic ci ies/ illages a e he ype o a ac ions p e e ed by highly ag eeable
pe sons, p obably he ones ha easily accompany amily and iends jus o make hem happy,
con i ming he esul s ound by Neidha d e al. (2015) and Jani (2014b). Con a y o ou p e ious
s udy, we could no ind ela ionship be ween he o he ou pe sonali y dimensions.
Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
123
Spo s & Games, like o enjoy spo ing compe i ions, play a he casino, and hun / ish, a e p e e -
ences p edic ed by high ex a e sion alues, which can de i e om he ene gy, exci emen and
g ega iousness inhe en o his ype o ac i i ies, bu also o he need o compe i i eness and dom-
ina ing/be he bes , con i ming he indings o Schneide and Vog (2012) and Neidha d e al.
(2015); bu nega i e ag eeableness, openness and conscien iousness, e i ying he indings o Jani
(2014b), who epo ed he Game ype was ela ed o low ag eeableness and conscien iousness
indi iduals. The di e ence in ag eeableness om he p e ious s udy can be due o he i ems e-
mo ed ela ed o p ac icing ball spo s and escape ooms, which a e ac i i ies ha in ol e coope a-
i eness and mo e open indi iduals. Nega i e openness and conscien iousness migh by ela ed o
indi iduals who a e willing o b eak ules and ac wi hou hinking.
Gas onomy ou s/ es i als a e posi i ely sough by ex a e ed and ag eeable indi iduals wi h
some nega i e openness and conscien iousness. This is in line wi h he esul s we p e iously ound
ha who enjoys wine and ood a e gene ally high-spi i ed and chee ul pe sons. A low conscien-
iousness and openness can be due o less awa eness o no ca ing o heal h issues pe sons,
gene ally ha ing a g ea pleasu e in ood/wine as ing and/o “addic ed” in consuming mo e han
needed allied o a low in ellec , as his ca ego y is no di ec ed o ine dining ac i i ies.
All pe sonali y dimensions a e ela ed o Heal h & Wellbeing (a ending SPA/beau y cen e s and
heal h and wellness ea men s), wi h a clea p edic ion by highly ex a e ed (a e no wo ied
abou exhibi ing hei body/in imacy), sligh ly conscien ious (ca e o hei heal h/wellbeing) and
neu o ic indi iduals (again wo ied o hei heal h, o ha s ess ou easily, being his so o ac i i-
ies a way o elaxing), wi h low openness and ag eeableness (mo e conse a i e, p e e ing ou ine
and mo e in e es ed in hei own p oblems).
Finally, a new ou ism ca ego y ela ed o obse ing Na u al Phenomena (like isi ing
ca es/ olcanoes o assis ing o no he n ligh s, olcanic e up ions) a ose. A posi i e ag eeableness
is he mos weigh ing dimension, ollowed by ex a e sion, and a nega i e conscien iousness,
which is he same p o ile as o Wild Na u e Ac i i ies.
Al hough some pe sonali y dimensions did no ha e a signi ican co ela ion o he choice o ce ain
ou is a ac ions, i does no mean hose co ela ions do no exis , bu ha a g ea e and mo e
ep esen a i e sample o each ype o p o ile is needed. Also, i is impo an o no e ha he e-
Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
124
sul s ound show ha only some cha ac e is ics om each pe sonali y dimension explain he p e -
e ences o ou is a ac ions, indica ing ha he p e e ences could be ine p edic ed by using a
ques ionnai e o e alua e each dimension’s six ai s mo e p ecisely, suppo ing he issues epo -
ed by Yee e al. (2011). Fo example, a pe son conside ed ex a e ed may no be a isk ake o
like ad enaline ac i i ies. I wouldn’ be e y good i he RS sugges ed bungee jumping o he ou -
is .
All he di e en ou ism ca ego iza ions and pe sonali y dimensions all sho when compa ed o
he indi idual expe ience ha a speci ic des ina ion/a ac ion can o e , o example, he medie al
ai in Sines, Po ugal, is one o he g ea es ai s, ec ea ing his o ical momen s, bu he e a e
o he s ha a e simple and a e mo e comme cially o ien ed.
5.4.2.3
Pe sonali y s T a elling Mo i a ions
As a esul o pe o ming he EFA on he 28 i ems ep esen ing he a elling mo i a ions, ou
i ems we e elimina ed
27
acco ding o he c i e ia p e iously e e ed (Sec ion 5.3), esul ing in a
inal scale wi h 24 i ems, and 6 ac o s ex ac ed ha explained 62% o he o al a iance. The 6
ac o s agg ega ed i ems measu ing he same concep s o somehow ela ed, as shown in Table
5.9 and by hei high C onbach’s Alpha eliabili y alues. The sampling adequacy (Kaise -Meye -
Olkin, KMO = 0.853) is good (Pes ana & Gagei o, 2008), and he co ela ion be ween he a iables
is signi ica i e (Ba le 's Tes o Sphe ici y Sig. = 0,000, < 0,05). The eliabili y o he ull scale is
good (
α
= 0.875), con i ming he i ems in he scale a e all ela ed o a elling mo i a ions, and
can he e o e be used as a e e ence. The ob ained ac o s we e named o meaning ul desc ip ions
ep esen ing he concep s we belie e hey symbolized.
As can be seen in Table 5.9, excep he ou i ems ha add o be emo ed and M12 and M26, all
pai s o i ems ha we e measu ing he same concep s in Pea ce and Lee (2005) scale agg ega ed
oge he a e he EFA. We can also see ha he pai s o i ems belonging o di e en dimensions in
Pea ce and Lee (2005) agg ega ed oge he in his s udy o cons i u e a new dimension, educing
he sample dimension, meaning he EFA conside ed hey we e measu ing simila concep s, which
27
Al hough agg ega ing in ac o s ha seemed o measu e he same concep s, he i ems wi h coe icien s < 0.40, ha we e sco ing alone, o ha we e
sco ing in mo e han one ac o , wi h a di e ence less han 0.10, we e elimina ed om he sample: M2, M7, M16, M21. Mo e de ails can be ound a
h p://www.gecad.isep.ipp.p /g ouplanne /dissemina ion.h ml
Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
125
is en o ced by he ob ained high eliabili y alues, excep o he Escape Obliga ions ac o , which
was on he limi o accep ance.
Table 5.9. Va imax o a ed componen ma ix o he p oposed T a elling Mo i a ions, showing he 6 ac o s ex ac ed using EFA
and hei espec i e i ems, he es ima ed co ela ions be ween he i ems and ac o s, and each ac o ’s C onbach’s Alpha elia-
bili y (α).
Fac o
I em
Desc ip ion
Es ima ed co ela ion
α
Sel -de elopmen &
Reliance
FM1
M20
Wo king on my pe sonal/spi i ual alues
0.723
0.828
M6
Unde s anding mo e abou mysel
0.715
M18
De eloping my skills and abili ies
0.685
M19
Being nea conside a e people
0.628
M5
Being wi h espec ul people
0.599
M4
De elop my pe sonal in e es s
0.523
Connec edness &
Recogni ion
FM2
M9
Ha ing oman ic ela ionships
0.717
0.784
M10
Showing o he s I can do i
0.712
M24
Being ecognized by o he people
0.688
M23
Mee ing amo ous pa ne s
0.676
M22
Re lec ing on pas memo ies
0.622
M8
Thinking abou good imes I’ e had in he pas
0.536
No el y & Exci emen
FM3
M25
Expe iencing some hing di e en
0.762
0.745
M17
Expe iencing h ills
0.699
M3
Ha ing da ing/ad en u esome expe iences
0.691
M11
Feeling he special a mosphe e o he aca ion des i-
na ion
0.557
Bond & Relax
FM4
M13
Doing some hing wi h my amily/ iend(s)
0.834
0.713
M27
S eng hening ela ionships wi h my amily/ iend(s)
0.733
M12
Ge ing away om e e yday psychological
s ess/p essu e
0.601
Na u e enjoymen
FM5
M1
Being close o na u e
0.858
0.845
M15
Ge ing a be e app ecia ion o na u e
0.824
Escape obliga ions
FM6
M14
Being obliga ed o no one
0.756
0.599
M26
Ge ing away om he usual demands o li e
0.609
M28
Being independen
0.567
Full scale α
0.875
The ull scale α is also p esen ed a he end o he able
When pe o ming he CFA, i ems M9 and M23, bo h ela ed o omance, we e emo ed om he
Connec edness & Recogni ion ac o due o a eg ession weigh < 0.50. We also decided o emo e
he Escape Obliga ions ac o , no only o ha ing a low eliabili y, bu because i had wo i ems
wi h a eg ession weigh < 0.50, esul ing in he model p esen ed in Figu e 5.5. The i e ac o s’
eg ession weigh s we e s a is ically signi ican in he p edic ion o hei espec i e i ems o p <
0.001*** ( wo- ailed). The model e ealed an o e all good goodness-o - i (
χ
2
/d =5.999,
CFI=0.912, GFI=0.921, PCFI=0.709, PGFI=0.645, RMSEA=0.070, p(RMSEA ≤ 0.05)=0.000),
meaning he model is alid o he s udy and he i ems p o ide a good i , con i ming he p oposed
“T a elling Mo i a ions” model. We can also obse e he model includes he mos common a el-
Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
126
ling mo i es ound in li e a u e: No el y & Exci emen (Explo a ion), Na u e enjoymen (Na u e ex-
pe iences), and Bond & Relax (Relaxa ion/Escapism).
The SEM o con i m wha pe sonali y dimensions we e p edic ing which a el mo i a ions (Figu e
5.6) e ealed an o e all accep able i (
χ
2
/d =6.379, CFI=0.822, GFI=0.849, PCFI=0.721,
PGFI=0.695, RMSEA=0.072, p(RMSEA ≤ 0.05)=0.000), con i ming he model is alid o he s udy
and he i ems p o ide an accep able i , con i ming he p oposed “Pe sonali y s T a elling Mo i a-
ions” model.
A a i s imp ession, i migh seem di icul o explain he obse ed co ela ions. We belie e he
s a is ically signi ican posi i e in luence o neu o icism on Sel -de elopmen & Reliance, Connec -
edness & Recogni ion, and No el y & Exci emen mo i a ions (Table 5.10) is ela ed o indi iduals
who a e sel -conscien (in e nal con ol locus) o he need o wo k on hose cha ac e is-
ics/capaci ies om an emo ional in elligence pe spec i e, ha ing he capabili y o looking a hem-
sel es, ecognizing he need o con ac wi h o he s and hemsel es, which can be good o open
ho izons and ha e di e en pe spec i es.
We can say openness o expe ience is somehow ela ed o all ou ism mo i a ions, being associa -
ed o expe iencing di e en hings, cu iosi y, ha ing a g ea e weigh on Sel -de elopmen & Reli-
ance and Bond & Relax, mo i a ions s ongly ela ed o indi iduals wi h a highe in ellec , ha need
o be s imula ed, p one o wande he mind o , and empa he ic o sel and o he s’ eelings. The
mo i a ions ela ed o openness a e suppo ed by Abba e and Di Nuo o (2013); Sca idi Abba e e
al. (2017) and Kashdan e al. (2009) s udies.
Cu iously, he ag eeableness, ex a e sion, and conscien iousness pe sonali y dimensions could
no p edic he p oposed a elling mo i a ions, ac ually hey had o be emo ed om he model
due o e y low eg ession weigh s in he co esponding i ems. As mo i a ions o a el depend on
many ac o s, such as he con ex , des ina ion, a ele ’s mood, ime o yea , companions, e c., i
can explain why no all dimensions could be ela ed o he mo i a ions o a elling (Table 5.10).

Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
127
5.4.2.4
Pe sonali y s T a el- ela ed P e e ences and Conce ns
As a esul o pe o ming he EFA on he 29 i ems ep esen ing he ele an a elled- ela ed p e -
e ences and conce ns o his s udy, se e al i ems we e elimina ed
28
acco ding o he c i e ia p e-
iously e e ed (Sec ion 5.3), esul ing in a inal scale wi h 18 i ems, and 4 ac o s ex ac ed ha
explained 50% o he o al a iance. The 4 ac o s agg ega ed i ems measu ing he same concep s,
as shown in Table 5.11 and by accep able C onbach’s Alpha eliabili y alues
29
. The sampling
adequacy (Kaise -Meye -Olkin, KMO = 0.805) is good (Pes ana & Gagei o, 2008), and he co ela-
ion be ween he a iables is signi ica i e (Ba le 's Tes o Sphe ici y Sig. = 0,000, < 0,05). The
eliabili y o he ull scale is accep able (
α
= 0.682), con i ming he i ems in he scale a e all ela -
ed o he same concep , and can he e o e be used as a e e ence. The ob ained ac o s we e hen
named o meaning ul desc ip ions ep esen ing he concep s we belie e hey symbolized.
The EFA on he i ems o he T a el- ela ed P e e ences and Conce ns e ealed in e es ing and no
so ob ious agg ega ions, showing many i ems we e measu ing he same concep s o somehow
ela ed. Rega ding conce ns, he pa icipan s conside impo an o eel sa e and com o able, a e
no willing o ake isks ega ding he physical in eg i y, and wan o ha e some so o p e isibili y
o a oid uncom o able o isky e en s, p e e ing he amilia o he unknown, which is easie
when a elling wi h amilia s o iends, ins ead o s ange s in a g oup. As o a el- ela ed p e -
e ences, o ha e cul u al expe iences, like isi ing he mos amous monumen s, lea n abou he
des ina ion’s his o y, see hings di e en om hei cul u e, and isi unusual/exo ic places, a e he
mos impo an o he esponden s. All hese ou comes suppo he esul s ound in li e a u e,
p e iously de ailed in Sec ion 5.2.2.3.
The CFA on he p oposed ac o s led o he emo al o he i ems P13, P15, P24R, and P28, as
hey had a eg ession weigh < 0.50, esul ing in he model p esen ed in Figu e 5.7. The ou ac-
o s’ eg ession weigh s we e s a is ically signi ican in he p edic ion o hei espec i e i ems o p
< 0.001*** ( wo- ailed). The model e ealed an o e all e y good goodness-o - i (
χ
2/d =2.754,
CFI=0.958, GFI=0.974, PCFI=0.737, PGFI=0.649, RMSEA=0.041, p(RMSEA ≤ 0.05)=0.982),
28
Al hough agg ega ing in ac o s ha seemed o measu e he same concep s, he i ems wi h coe icien s < 0.40, communali ies < 0.50, ha belonged o
ac o s wi h a low eliabili y (α < 0.60), ha we e sco ing alone, o ha we e sco ing in mo e han one ac o , wi h a di e ence less han 0.10, we e elimi-
na ed om he sample: P1, P3, P8, P9, P11, P16, P17, P20, P22, P23, P26. Please consul h p://www.gecad.isep.ipp.p /g ouplanne /
dissemina ion.h ml o u he de ails.
29
Being an explo a o y analysis, i is accep able o conside ac o s wi h C onbach’s Alpha alues ≥ 0.60 (Hai e al., 2009).
Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
128
meaning he model is alid o he s udy and he i ems p o ide a e y good i , con i ming he p o-
posed “T a el- ela ed P e e ences and Conce ns” model.
Figu e 5.5. Simpli ied S uc u al Equa ion Model o he p oposed “T a elling Mo i a ions” model, ob ained using CFA. The
i e ac o s’ eg ession weigh s we e s a is ically signi ican in he p edic ion o hei espec i e i ems o p < 0.001*** ( wo-
ailed)
Table 5.10. S anda dized eg ession weigh s o he ela ionship be ween he BFI dimensions and T a elling Mo i a ions, ob-
ained using CFA.
Fac o
BFI Dimension
Reg ession Weigh
p
Sel -de elopmen & Reliance FM1
Neu o icism
0.125
0.001**
Openness
0.429
***
Connec edness & Recogni ion FM2
Neu o icism
0.169
***
Openness
0.226
***
No el y & Exci emen FM3
Neu o icism
0.156
***
Openness
0.189
***
Bond & Relax FM4
Neu o icism
0.049
0.205
Openness
0.372
***
Na u e enjoymen FM5
Neu o icism
0.028
0.450
Openness
0.285
***
S a is ically signi ican alues a e in bold and i alics (** p < 0.01 (2- ailed), *** p < 0.001 (2- ailed))
Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
129
Figu e 5.6. Simpli ied S uc u al Equa ion Model o he p oposed “Pe sonali y s T a elling Mo i a ions” model, ob ained
using CFA. Fo eadabili y, he e o a iables ha e been emo ed. The i ems label can be ound a Table 5.6 and Table 5.9
The SEM o con i m wha pe sonali y dimensions we e p edic ing which a el- ela ed p e e ences
an conce ns (Figu e 5.8) e ealed an o e all accep able i (
χ
2/d =4.488, CFI=0.809, GFI=0.877,
PCFI=0.716, PGFI=0.733, RMSEA=0.058, p(RMSEA ≤ 0.05)=0.000), con i ming he model is
alid o he s udy and he i ems p o ide an accep able i , con i ming he p oposed “Pe sonali y s
T a el- ela ed P e e ences and Conce ns” model.
All pe sonali y dimensions a e ela ed o a el- ela ed p e e ences and conce ns, al hough wi h
lowe eg ession weigh s compa ed o he ou ism ca ego ies and a elling mo i a ions (Table
5.12). Indi iduals wi h a low ex a e sion and neu o icism, wi h some deg ee o ag eeableness and
openness, a e mo e p eoccupied wi h com o , sa e y and heal h conce ns (P e isibili y & Sa e y),
which can be con i med by he indings o Ma i z e al. (2013), Tan (2020), Ca alho e al. (2020),
and Al-Omi i e al. (2021). The conscien iousness eg ession weigh was no s a is ically signi ican
o be conside ed.
Chap e 5 (Pape 3) - G oup ecommende sys ems o ou ism: how does pe sonali y p edic p e e ences o a ac ions, …
130
Table 5.11. Va imax o a ed componen ma ix o he p oposed T a el- ela ed P e e ences and Conce ns, showing he 4 ac o s
ex ac ed using EFA and hei espec i e i ems, he es ima ed co ela ions be ween he i ems and ac o s, and each ac o ’s
C onbach’s Alpha eliabili y (α).
Fac o
I em
Desc ip ion
Es ima ed
co ela ion
α
P e isibili y &
Sa e y
FP1
P14
Fo me, o eel com o is always he mos impo an
(quali y o acili ies / p oduc s)
0.700
0.731
P6
I would ne e a el o a place whe e he e was no
mobile phone ne wo k
0.690
P12
In a dis an coun y, one o my wo s ea s would be o
ge los
0.682
P15
Fo me, o ul ill expec a ions is mo e impo an han a
good su p ise
0.564
P5
I am a aid o ge ing ill o ha ing acciden s whils away
on aca ions
0.551
P2
Unde no ci cums ances I like o ake isks ela ed o
my physical in eg i y
0.531
P28
I would be incapable o a elling o a high c iminali y
a e / a med con lic des ina ion
0.504
Cul u al &
Lea ning
Expe iences
FP2
P18
Fo me, a good aca ion mus include a cul u al / lea n-
ing componen
0.793
0.782
P4
When going on aca ions I ake in o accoun he des i-
na ion's cul u al o e
0.783
P10
Be o e a elling I like o know/s udy he his o y o he
des ina ion
0.703
P19
I would ne e isi a g ea ci y wi hou seeing i s iconic
monumen s
0.671
P7
When planning aca ions, I y o include as many plac-
es/a ac ions as possible
0.607
Uniqueness &
Exo icness
FP3
P21
Fo me, i is impo an o see exo ic hings o ha a e
e y di e en om my cul u e
0.729
0.615
P29
I like o isi uncommon places o obse e peculia
hings (e.g., wo ld eco ds, pop icons, his o ical i ems,
e c.)
0.688
P13
I like o go whe e ew people ha e been o be o e
0.679
Familia i y
FP4
P27
I would ne e a el in a g oup due o p i acy easons
0.777
0.608
P24R
I would be willing o a el in a g oup o ganized by a
a el agency
0.772
P25
I would ne e go on aca ion wi h s ange s (making
common ips and meals)
0.659
Full scale α
0.682
The ull scale α is also p esen ed a he end o he able. R deno es e e sed ques ions.
The conce n o a elling in g oups o s ange s (Familia i y) is p edic ed by a nega i e ag eeable-
ness ( e ealing indi iduals mo e conce ned abou hemsel es), neu o ic (who a e mo e com o a-
ble wi h amily/ iends ha ing di icul y in socializing wi h s ange s) and conscien ious pe sons
(ca e ul and less spon aneous). To he bes o ou knowledge, we could no ind wo ks ela ing he
Big Fi e o his ype o conce ns. The o he dimensions we e no s a is ically signi ican .
Chap e 9 (Pape 7) - "You Wan o Play a Game?" De ec ing Two Pe sonali y T ai s wi h Sho -Du a ion Mobile Games
233
o he social desi abili y bias (Ped egon e al., 2012) and ake esponses (Bha ia & Ryan, 2018).
Also, hey a e mainly ocused in b oade pe sonali y dimensions ins ead o he mo e g anula pe -
sonali y ai s, which a e be e o cha ac e izing a pe son (Cos a & MacC ae, 1992), and he ones
who in end o de e mine he mo e g anula ai s a e no ha accu a e (F. Y. Wu e al., 2022). The
challenge is he e o e o accu a ely ga he he use s’ pe sonali y non-in usi ely and in a less ime-
consuming and mo e compelling way, wi hou he need o machine lea ning (ML) echniques,
which need g ea amoun s o da a and use in e ac ions o p o ide accu a e p edic ions.
An applica ion example a e Recommende Sys ems (RS) o ou ism. T a elling is an emo ional
expe ience (Volo, 2021) and he e o e, pe sonaliza ion is a key ac o o he success o RS in ou -
ism (Ga alas & Ken e is, 2011; Ricci, 2002; Schmid -Belz e al., 2002). The mo e in o ma ion
abou he ou is is known, he be e ecommenda ions can be made. Fo ins ance, pe sonali y is
s ongly ela ed o he use s’ p e e ences (Al es, Ma ins, Sa ai a, e al., 2023; Can ado e al.,
2013; F iedman & Schus ack, 1999; Tkalcic & Chen, 2015) and has been e idenced o imp o e
ecommenda ions and ackle he cold-s a p oblem (Al es, Ma ins, e al., 2024; Feil e al., 2016;
Tkalcic & Chen, 2015). The e o e, he use s’ pe sonali y can help build a mo e obus p o ile, and
i somehow needs o be ob ained by he sys em.
Gami ica ion and se ious games can be he le e age we a e looking o . In ac , gami ica ion ech-
niques will be a majo end o he u u e o ou ism (Mo a a e al., 2014; Xu e al., 2017; Xu e
al., 2016). Fo example, Dunn e al. (2009) compa ed h ee di e en in e aces o acqui ing pe -
sonali y in RS: explici ly, using he NEO PI-R pe sonali y ques ionnai e, and implici ly, wi h he
Commons Fishing (compu e ) Game and he Implici Associa ion Tes . They ound he game was a
good p edic o o he Ag eeableness and Ex a e sion dimensions bu needed o be imp o ed o
accu acy. They said i would be p omising o explo e o he game ( heo y) al e na i es ha could be
applied o RS. Thus, adding se ious games o p o ile use s can be he wis needed in he his o y o
RS and o o he pe sonali y-based sys ems/applica ions.
Se e al a emp s o p o iling playe s ha e been made (Diamond, Tondello, Ma czewski, Nacke, &
Tscheligi, 2015; Halim, A i , Rashid, & Edwin, 2017), s a ing wi h Ba le (1996), who p oposed a
playe ypology o mul iplaye games composed o ou playe ypes: achie e s, explo e s, social-
ize s and kille s. Howe e , in his wo k, we in end o p o ile playe s based on hei c ude (and mo e

Chap e 9 (Pape 7) - "You Wan o Play a Game?" De ec ing Two Pe sonali y T ai s wi h Sho -Du a ion Mobile Games
234
g anula ) pe sonali y ai s, using he mos used and bes ecognized pe sonali y model, he Big
Fi e (Digman, 1990), so hei p o ile can be used ac oss mul iple domains.
Thus, we p opose o implici ly acqui e he indi idual’s mo e g anula Big Fi e pe sonali y ai s in
less han i e minu es, by using sho -du a ion mobile games, o mo i a e and challenge he use s
a he same ime, as a means o subs i u ing he sel - epo ing pe sonali y ques ionnai es, s a ing
wi h wo ai s as a i s p oo o concep : cau iousness and achie emen -s i ing. Two 2D pla o m
minigames we e de eloped, one mo e ocused in he de ec ion o he cau iousness ai , and he
o he in he achie emen -s i ing ai . An expe imen wi h eal use s (
𝑛
= 100) om di e en ages,
o ma ion a eas, and p o essions was conduc ed. The In e na ional Pe sonali y I em Pool (IPIP)
IPIP-NEO-120 pe sonali y ques ionnai e was hen used o assess he pa icipan s’ pe sonali y
g anula ai s and compa ed wi h he minigames esul s. The achie emen -s i ing ai could be
success ully associa ed wi h game beha io s, bu cau iousness could no , al hough i seemed o be
p omising wi h he igh imp o emen s and a la ge sample. S a is ically signi ican co ela ions
wi h o he pe sonali y ai s we e also ound, namely: ange , modes y, iendliness, exci emen
seeking, chee ulness, ad en u ousness, anxie y, imagina ion, in ellec , sel -conscien iousness, and
ulne abili y.
In Sec ion 9.2 we in oduce he concep o pe sonali y and i s ela ionship o he use s’ p e e -
ences, and he di e en echniques used o acqui e he use s’ pe sonali y, including se ious
games, and some ela ed wo ks. Sec ion 9.3 exposes he used me hodology. Sec ion 9.4 shows
he ob ained esul s and espec i e analysis, and inally, Sec ion 9.5 e lec s on he s udy esul s
and desc ibes wha will be done as u u e wo k.
9.2
B
ACKGROUND AND
R
ELATED
W
ORK
9.2.1
Pe sonali y and Use P e e ences
Psychological aspec s, such as pe sonali y, moods and emo ions, a e being pe cei ed o in luence
he a iance in he indi idual’s p e e ences and beha io (Dhelim e al., 2021; T. T. Nguyen,
Maxwell Ha pe , Te een, & Kons an, 2018; Tkalčič e al., 2016), and hei conside a ion is e i-
dencing o show be e esul s han gene ic app oaches (Cas ells & Jannach, 2024; Nunes e al.,
2008; Tondello e al., 2017; Wa is e al., 2024).
Chap e 9 (Pape 7) - "You Wan o Play a Game?" De ec ing Two Pe sonali y T ai s wi h Sho -Du a ion Mobile Games
235
As de ined by H. Eysenck and Rein (1998), “pe sonali y is he sum- o al o he ac ual o po en ial
beha io -pa e ns o he o ganism, as de e mined by he edi y and en i onmen ”. Each pe son has
he own beha io pa e ns, which a e a he s able o e ime in di e en si ua ions (McC ae &
Cos a J , 1997). These pa e ns can be summa ized in o pe sonali y dimensions (o ac o s), being
he Fi e Fac o Model, OCEAN, o Big Fi e, ecognized as he bes model o ep esen hem
(Digman, 1990): Openness o Expe ience (O), Conscien iousness (C), Ex a e sion (E), Ag eeable-
ness (A), and Emo ional S abili y/Neu o icism (N). Each ac o is u he e ined in o six ace s o
ai s
48
(Cos a & MacC ae, 1992), which a e mo e g anula and can be used o be e po ay a
pe son. The hi y ai s a e p esen ed in Table 9
.1
, wi h he s udied wo ai s in bold.
Table 9.1. Big Fi e Pe sonali y dimensions and Co esponding six ai s, Adap ed F om Cos a and MacC ae (1992)
.
Openness
Conscien iousness
Ex a e sion
Ag eeableness
Neu o icism
Imagina ion
Sel -e icacy
F iendliness
T us
Anxie y
A is ic in e es s
O de liness
G ega iousness
Mo ali y
Ange
Emo ionali y
Du i ulness
Asse i eness
Al uism
Dep ession
Ad en u ousness
Achie emen -s i ing
Ac i i y le el
Coope a ion
Sel -consciousness
In ellec
Sel -discipline
Exci emen -
seeking
Modes y
Immode a ion
Libe alism
Cau iousness
Chee ulness
Sympa hy
Vulne abili y
In bold he wo s udied pe sonali y ai s.
Fo ins ance, as no ed by Tkalcic and Chen (2015), pe sonali y can be use ul in many a eas o RS,
since i is s ongly connec ed o he use s’ choices (Ma ijn e al., 2022). Use s wi h simila pe -
sonali ies ha e a g ea endency o choose simila con en s o i ems. Fo example, in ou ism, high-
ly ex a e ed and low conscien iousness indi iduals end o ake mo e isks and p e e ad enaline
ac i i ies (Al es, Ma ins, Sa ai a, e al., 2023). In he music ield, popula music is enden ially
sough by ex a e ed pe sons who a e in need o posi i e emo ions, g ega iousness, and wa m h,
and ock music is enden ially enjoyed by pe sons who sco e high on exci emen seeking (Rawlings
& Cianca elli, 1997). E en he cha ac e is ics o use s’ pic u es on Ins ag am a e ela ed o pe -
sonali y dimensions (Fe we da e al., 2015). The e o e, pe sonali y is a powe ul ai in humans
ha can be le e aged o p edic hei p e e ences ac oss a ious a eas.
In his wo k, we p oposed o de ec wo pe sonali y ai s om he Conscien iousness dimension o
he Big Fi e, cau iousness and achie emen -s i ing, h ough wo minigames de eloped o And oid
mobile de ices. The choice o hese ai s was mainly because, based on ou expe ience wi h ide-
48
Fo simplici y, om now on called ai s.
Chap e 9 (Pape 7) - "You Wan o Play a Game?" De ec ing Two Pe sonali y T ai s wi h Sho -Du a ion Mobile Games
236
ogames, he e a e se e al popula ideogames ha seem easie o “mimic”, ha ing ea u es ha
we hink could help measu e hose ai s, and, o he bes o ou knowledge, he e a e a ew s ud-
ies ha y o implici ly acqui e he playe s’ cau iousness and achie emen -s i ing wi h se ious
games, bu wi hou much success (see Sec ion 9.2.2).
We can desc ibe someone cau ious as p uden , hough ul, and delibe a e in he decisions, making
plans ca e ully (Cos a J e al., 1995). Cau ious indi iduals p e e o a oid unnecessa y isks, a e
a en i e o de ails, p e e well-es ablished ou ines, and may lack spon anei y. This ai is associ-
a ed wi h a desi e o secu i y. Pe sons wi h a high le el o achie emen -s i ing a e o en ambi-
ious, pe sis en , and dedica ed o hei pu sui s (Cos a J e al., 1995; Ellio & McG ego , 2001).
They se demanding s anda ds o hemsel es and ac i ely seek success and accomplishmen in
a ious aspec s o hei li es. Indi iduals wi h his ai a e ypically ocused on con inuous im-
p o emen , a e willing o pu in e o and ha d wo k, being o en “wo kaholic”, and a e d i en by a
desi e o a ain excellence. The opposi e can be said o indi iduals low on he e e ed ai s.
9.2.2
Pe sonali y Acquisi ion
The e a e wo ways o acqui ing he use ’s pe sonali y (Tkalcic & Chen, 2015): by using (1) explici
echniques – usually by means o Psychology ques ionnai es
49
; o (2) implici echniques: o ex-
ample by da a mining emails o social media like Facebook.
An inc easing numbe o s udies show ha pe sonali y dimensions can be ecognized om se e al
human cues and ha di e en kinds o (combina ion o ) me hods can be used o au oma ically
acqui e hem (Finne y e al., 2016; Mohammadi, Vincia elli, & Mo illa o, 2010; K. A. Smi h,
Dennis, Mas ho , & Tin a e , 2019; Tkalcic & Chen, 2015). Finne y e al. (2016) compiled i e
di e en app oaches ha can be used o implici ly ecognize someone’s pe sonali y:
1.
Audio and/o isual app oaches: Cues like a pe son’s speaking a e, eye gaze, acial ex-
p essions, pos u es, ges u es, oice one and pi ch, among o he s, can ex e io ize pe son-
ali y dimensions. Fo ins ance, ex o e ed pe sons end o speak mo e, as e and hesi a e
less han in o e ed (Sche e , 1979). Inspi ed in he HCRC Map Task o Ande son e al.
(1991), Ba inca, Lep i, Mana, and Pianesi (2012) ecognized ou pe sonali y dimensions
49
To measu e pe sonali y se e al i ems and scales exis and can be consul ed a he IPIP websi e a h ps://ipip.o i.o g/
Chap e 9 (Pape 7) - "You Wan o Play a Game?" De ec ing Two Pe sonali y T ai s wi h Sho -Du a ion Mobile Games
237
om he in e ac ion o es pa icipan s wi h a map in a compu e ( isual and acous ic
cues), by gi ing e bal ins uc ions ia Skype. They used he Ab idged Big Fi e-dimensional
Ci cumplex model pe sonali y ques ionnai e as a baseline o di ec ly measu e he pe sonal-
i y dimensions. They ound Ex a e sion and Emo ional S abili y we e he easies dimen-
sions o de ec , ollowed by Ag eeableness and Conscien iousness.
2.
Wea able de ices: Da a can be eco ded om wea able de ices wi h senso s, like sma -
wa ches, elec onically ac i a ed eco de s o sociome ic badges, wi hou in e up ing he
use ’s daily li e (Finne y e al., 2016). Olguın, Gloo , and Pen land (2009) showed ha i is
possible o iden i y pe sonali y dimensions by using a sociome ic badge. They ins umen -
ed 67 nu ses in a Hospi al om Bos on and eco ded hei ac i i ies while hey wo ked. 39
pa icipan s answe ed he NEO-FFI pe sonali y ques ionnai e (60 i ems). Fo ins ance, hey
ound ha nu ses who spen mo e ime o he day on a ace- o- ace in e ac ion, and ha
had mo e a ia ion ac oss days in he daily pe cen age o ace- o- ace in e ac ion ime, he
mo e neu o ic hey we e.
3.
Sma phones: We can say sma phones a e a kind o po able mul i-senso de ice almos
e e yone ca ies. Due o hei inc easing p ocessing and memo y capaci ies, and numbe
o buil -in senso s, hey can be used o eco d and sense a panoply o da a ela ed o a us-
e ’s daily li e, like his p ecise loca ion, o he de ices in his icini y, communica ion da a
and hei con en (e.g. phone calls logs, ex messages, emails), ime o day, wea he ,
mood, o e en mo emen pa e ns. Fo ins ance, i is e idenced ha ex o e s end o e-
cei e mo e calls and spend mo e ime on hem (Chi a anjan, Blom, & Ga ica-Pe ez,
2013).
4.
Tex : W i en ex s can be used o p edic pe sonali y by using ex -mining echniques. Tex-
ing me ges o al and w i en communica ion and can be o ex eme in e es (Finne y e al.,
2016). Fo example, ex a e s ex mo e equen ly and use mo e pe sonal p onouns
(Hol g a es, 2011).
5.
Social Media: Social media like Facebook, Twi e , o Ins ag am, can also be a means o
showing someone’s pe sonali y. Se e al cha ac e is ics can be ex ac ed om he use s’
p o iles, like he numbe o ollowe s, he ne wo k’s densi y, numbe o pos s o links (Gol-
Chap e 9 (Pape 7) - "You Wan o Play a Game?" De ec ing Two Pe sonali y T ai s wi h Sho -Du a ion Mobile Games
238
beck, Robles, Edmondson, & Tu ne , 2011), numbe o likes, numbe and ype o sha es,
numbe o g oups (Kosinski, Bach ach, Kohli, S illwell, & G aepel, 2014), e c. Many social
con ac s (popula use s) can be a signal o ex a e sion and less neu o icism (Golbeck e
al., 2011). Celli, B uni, and Lep i (2014) showed ha e en p o ile pic u es om Facebook
can e eal a lo o in o ma ion abou a use . Fo ins ance, hose who a e “no open o ex-
pe ience end o ha e s ong spo s o ligh and some imes da kened igu es. A ec i e peo-
ple end o smile in p o ile pic u es. Dominan people end o ha e e y b igh , isible o a -
ac i e pic u es” (Celli e al., 2014). The au ho s used he BFI-10 ques ionnai e o meas-
u e he pa icipan s’ pe sonali y and he IPIP In e pe sonal Ci cumplex IPIC-IPC-32 o as-
sess hei in e ac ion s yles.
6.
(Se ious) Games:
I has also been ound ha a pe son’s beha io in (compu e ) games e-
eals hei pe sonali y (Cowley & Cha les, 2016; Van Lank eld e al., 2011; Vincia elli &
Mohammadi, 2014; Yee e al., 2011). The s udy o he playe s s a ic beha io al pa e ns,
such as pe sonali y, in he con ex o playing a compu e game, is called playe p o iling
(Yannakakis, Sp onck, Loiacono, & And é, 2013). Al hough Finne y e al. (2016) do no e-
e his app oach in hei su ey, i is e e ed by o he au ho s (Tkalcic & Chen, 2015;
Vincia elli & Mohammadi, 2014). So, we decided o include i in his sec ion.
Fo ins ance, Van Lank eld e al. (2011) made a e y in e es ing s udy by c ea ing a mod-
ule o he Massi ely Mul iplaye Online Role-Playing Game (MMORPG) Ne e win e Nigh s
o p edic he game s’ pe sonali y. They concluded ha all he i e pe sonali y dimensions
could be p edic ed om he pa icipan s’ in e ac ions wi h he game s o yline. They also e -
idenced ha each dimension can be be e p edic ed i si ua ions op imally sui ed o each
dimension a e inse ed in he game. They used he NEO-PI-R pe sonali y ques ionnai e o
assess he pa icipan s’ pe sonali y dimensions.
Yee e al. (2011) used he beha io s om 1040 playe s in Wo d o Wa c a , also a
MMORPG, o de e mine i hey we e ela ed o hei pe sonali y. Fo ou mon hs, hey col-
lec ed he playe s da a and compa ed he esul s wi h he pa icipan s’ Big Fi e pe sonali y
dimensions, ob ained om a 20-i em scale e ie ed om he IPIP. They ound many be-
ha io al cues we e ela ed o he playe s’ pe sonali y and he e o e could be used o implic-

Chap e 9 (Pape 7) - "You Wan o Play a Game?" De ec ing Two Pe sonali y T ai s wi h Sho -Du a ion Mobile Games
239
i ly de ec i . Fo ins ance, g oup ac i i ies we e p e e ed by ex a e ed pe sons and solo
by in o e ed. High openness o expe ience indi iduals had mo e cha ac e s and a highe
in e es in explo ing he su oundings, he less conscien ious seemed o be less ca e ul and
died mo e o en om alls om high loca ions.
San os, Hu chinson, Khan, and Ma kopoulos (2019) ied o de ec wo abs ac pe sonali y
ai s, need o cogni ion and sel -es eem, h ough he playe s’ game beha io , in a simple
s a egy game and in a 2D pla o me , espec i ely. They disco e ed ha playe s wi h low
need o cogni ion used mo e hin s in he game han high need o cogni ion playe s. Sel -
es eem seemed o be ela ed o he playe s’ pace in he game.
Palomäki e al. (2021) s udied how ex a e sion and conscien iousness we e ela ed o
online ace ho se be ing in a sample o 11217 Finnish males online ho se be o s. They
ound high ex a e sion indi iduals we e mo e p one o being be o s, placing mo e be s
pe day, and wi h a g ea e be ing olume. High conscien ious pe sons (mo e cau ious)
we e less likely o be be o s, placing less be s pe day, wi h less money, and playing ewe
days.
Click Town and Wo d Find a e wo game-based assessmen s de eloped by F. Y. Wu e al.
(2022) in an a emp o measu e h ee ai s o Conscien iousness: achie emen -s i ing,
sel -discipline, and cau iousness. They used he 50-i em NEO-PI-R o assess he pa ici-
pan s’ pe sonali y. Howe e , hey ound he games we e measu ing he pa icipan s cogni-
i e abili y ins ead o he in ended ai s, gi ing se e al ecommenda ions o u u e e-
sea ch.
Quwaide e al. (2023) de eloped and published in he Google Play S o e a mobile i s -
pe son shoo e game, “The P o ec o ”, whe e he main goal is o kill zombies o sa e se -
e al ci ies, o p edic he playe s’ Big Fi e pe sonali y dimensions. To de e mine each di-
mension, hey used h ee di e en cus omizable scena ios, each o ep esen one ai o
he co esponding dimension, being one o hem he achie emen -s i ing ai . They used
he accu acy o zombies killed as a me ic o measu e ha speci ic ai . They also com-
pa ed se e al machine lea ning me hods o p edic he playe s’ pe sonali y. Howe e , how
he pa icipan s pe sonali y was ob ained and compa ed o he measu ed me ics is no e-
Chap e 9 (Pape 7) - "You Wan o Play a Game?" De ec ing Two Pe sonali y T ai s wi h Sho -Du a ion Mobile Games
240
ealed, and he e o e, how he achie emen -s i ing me ic in he game is co ela ed o he
pa icipan s’ co esponding ai is no known, and hus canno be used o compa ison.
The use o explici echniques o acqui e he use s’ pe sonali y is in usi e, ime-consuming, and
no e e yone is willing o answe a se ies o in ima e ques ions, being subjec o he well-known
desi abili y bias and ake esponses, which gene ally p ejudice sys ems ha ely on a ques ionnai e
o acqui e pe sonali y. To ackle his limi a ion, esea che s ha e s a ed o wo k on ways o ga he -
ing he use s’ pe sonali y in implici ways, like using se ious games. Se ious games a e becoming
popula and inc easingly being used o educa ional and aining pu poses in a g ea a ie y o
domains (A i e al., 2024; de Ca alho, B az, dos San os, Fe ei a, & P a es, 2024; Djaou i e al.,
2011; Susi e al., 2007; Xu e al., 2017).
To assess and alida e hei esul s, esea che s equen ly use a pe sonali y ques ionnai e, usually
ex ensi e, o ob ain “g ound u h” da a and hen co ela e hose alues wi h he da a ex ac ed by
he me hod hey wan o s udy, usually applying co ela ion o eg ession analysis. Depending on
he s udy pu pose, he da a ob ained can be u he subjec ed o ML echniques o make p edic-
ions. Howe e , ML echniques p esen limi a ions: hey need a g ea numbe o in e ac ions and
inpu om he use o ob ain enough da a o de e mine he use s’ beha io , o he s a e no e y
p ac ical, and hey a e no e y accu a e ye . Mo eo e , social media da a is being ques ioned
abou i s eliabili y in e ms o pe sonali y cues, since he in o ma ion p esen ed is selec ed by he
use and he e o e, i ep esen s wha he use wan s o e eal o he o he s and no he ue sel
(Finne y e al., 2016; Gosling, Vazi e, S i as a a, & John, 2004). Also, no e e yone in e ac s in
he same way wi h a social ne wo k: some a e me e obse e s, o he s only “like” ce ain
pos s/i ems, which poses a p oblem o app oaches ha need ha knowledge o make p edic-
ions.
Ano he limi a ion ound is ha mos o he exis ing s udies a e based on he b oade i e dimen-
sions o pe sonali y o on o he pe sonali y scales, and no on he mo e g anula hi y ai s. We
belie e he hi y Big Fi e’s ai s should be assessed ins ead i we wan o p edic mo e accu a ely
he use s’ pe sonali y, which can hen be used ac oss a a ie y o domains. Ano he limi a ion ha
s ands ou is ha he pe sonali y ai s a e ea ed sepa a ely as i he e was no co ela ion be-
ween hem (Finne y e al., 2016), which is no always he case (Vincia elli & Mohammadi, 2014).
Chap e 9 (Pape 7) - "You Wan o Play a Game?" De ec ing Two Pe sonali y T ai s wi h Sho -Du a ion Mobile Games
241
Fo example, inside he same pe sonali y dimension Conscien iousness, someone who is cau ious
migh no be ambi ious, and ice- e sa. So, i is impo an o keep in mind ha ela ionship and
ha i may be complex, o no e en possible, o measu e jus a single pe sonali y ai in a game,
as o he ai s a e ce ainly ela ed (posi i ely o nega i ely).
To ind a way o p edic ing someone’s pe sonali y wi h enough knowledge wi hou he need o a
g ea numbe o in e ac ions and con inuous inpu s om he use , in an unob usi e manne , is
he e o e a challenge. We p opose o o e come hose limi a ions by using sho -du a ion mobile
se ious games. Since mobile de ices can be aken anywhe e and (supposedly) be used any ime,
hey a e a pe ec and p ac ical means o acqui ing he use s’ pe sonali y and o he use -cen ic
in o ma ion. In sum, wi h his s udy, we in end o in oduce sho -du a ion mobile games o implic-
i ly de ec he use s’ g anula pe sonali y ai s, so hey can be gene alized o any ype o domain,
such as ou ism, educa ional, heal hca e, co po a e, among o he s, and in eg a ed in o pe sonali y-
based sys ems.
So, do you wan o play a game?
9.3
M
ETHODOLOGY
9.3.1
The Pe sonali y Minigames Design
To y o de e mine he playe s’ cau iousness and achie emen -s i ing, wo minigames, Which
Way (WW) and Time T a el Mania (TT), espec i ely, planned o ake no mo e han 5 minu es each
o comple e, we e de eloped using Uni y
50
. Bo h a e 2D pla o m games and ha e he same main
cha ac e , Be y, ha he playe mus con ol h ough he map, using he same ype o game con-
ols, o a oid a di e en lea ning cu e be ween he games. Bo h games ha e an ene gy ba ,
which ep esen s how much li e Be y has and a coun e o gold coins, as well as backg ound mu-
sic o y o make he game mo e imme si e and enjoyable o play. WW game has a black key icon
ha u ns yellow when he playe ca ches i , opening he end doo , and TT game has a diamonds
coun e and a ime in seconds (see
50
The games APK can be downloaded in he supplemen a y ma e ial (APK olde ). Please accep he wa nings in o de o ins all.
Chap e 9 (Pape 7) - "You Wan o Play a Game?" De ec ing Two Pe sonali y T ai s wi h Sho -Du a ion Mobile Games
242
Figu e 9.1). Con a y o o he wo ks ha used games o de e mine pe sonali y, he p oposed
games aimed o de ec pe sonali y ai s a he i s in e ac ion, wi hou needing o be played se -
e al imes o build g ound u h da a. The cha ac e is ics o he wo pe sonali y ai s (desc ibed in
Sec ion 9.2.1) se ed as basis o he easoning behind he games design.
The da a collec ed in bo h games was eco ded in a CSV ile a he end o each ull game a emp ,
on he playe s’ mobile phone, and is de ailed in Sec ion 9.3.6.
We ied o make he minigames easy o play o any age g oup, so ha hey we e no skill de-
penden o a oid in oducing a skill bias. Howe e , due o he ai s ha we e being measu ed,
some di icul ies had o be in oduced. The minigames a e de ailed nex .
Figu e 9.1. (Le ) Ini ial scene o he cau iousness pe sonali y minigame. (Righ ) Achie emen -s i ing minigame ini ial
sc een.
9.3.1.1
Cau iousness Minigame
In his game, he goal was o inish he game by inding a hidden key o unlock he inal doo . As
cau ious pe sons a e associa ed wi h isk-a e sion and low impulsi i y in di e en li e-s yle choices
(Palomäki e al., 2021), being conside ed p uden and ca e ul (Cos a J e al., 1995), eigh special
moni o ing sec ions (MS) we e designed o y o measu e ha isk a e sion and ca e ulness. They
consis ed o bi u ca ions, one o hem leading o an easy pa h and he o he o a ha d pa h. The
easy pa hs we e as , s aigh o wa d, wi h no obs acles and no isk o dea h, usually wi h one o
wo easy o ca ch gold coins. This could a ac mo e cau ious people. The ha d pa hs had a leas
he double o gold coins o gi e an ex a mo i e o ollow ha pa h, and ha d obs acles o pass
h ough, like spikes, wa e , og, o la a, which became ha de as he game p og essed, wi h an
ele a ed dea h isk. The spikes and la a killed ins an ly. The di e ence be ween he pa hs and he
chance o die on he ha d pa hs was in en ional, gi ing an ex a ewa d o isking he ha d pa h,
o he wise e e yone could choose he easies . A he beginning o he game, he playe could go
Chap e 9 (Pape 7) - "You Wan o Play a Game?" De ec ing Two Pe sonali y T ai s wi h Sho -Du a ion Mobile Games
249
old he WW game had bi u ca ions, leading o an easy o ha d pa h, and ha a e choosing a
pa h, he o he pa h would ge blocked. The exi key, checkpoin s and li e egene a ion we e also
explained. The playe s we e conside ed o make a comple e a emp only i hey eached he end
o he game. Be o e s a ing he expe imen s, he pa icipan s we e gi en a minu e o amilia ize
hemsel es wi h each game.
All he pa icipan s had o ill a p e-ques ionnai e, c ea ed using Mic oso Fo ms, composed o 9
ques ions ela ed o hei demog aphics and 3 ela ed o ideogames playing equency and pla -
o ms; a pos -ques ionnai e o each minigame, o ga he hei eedback on he games played,
including gameplay and asks pe o med; ollowed by he IPIP-NEO-120 pe sonali y ques ionnai e
(Johnson, 2014) o explici ly ga he hei pe sonali y ai s and compa e o he collec ed esul s.
This pe sonali y ques ionnai e is composed o 120 ques ions, using a 5-poin Like scale, o as-
sess an indi idual in he 30 ai s o he Big Fi e dimensions o pe sonali y o Lewis R Goldbe g
(1990). This ques ionnai e is a sho e e sion o he IPIP-NEO (300-i em), which is p o en o be
eliable and alid o de e mine he i e pe sonali y dimensions and he co esponding 30 ai s
(Johnson, 2014). The expe imen asks o de was he ollowing: 1) ill p e-ques ionnai e, 2) play
TT, 3) ill TT pos -ques ionnai e, 4) play WW, 5) ill WW pos -ques ionnai e, 6) ill IPIP-NEO-120
pe sonali y ques ionnai e.
Pape shee s wi h QR codes we e gi en o he pa icipan s o download he games APK and access
he ques ionnai es.
9.3.6
Measu emen s
The ollowing da a was collec ed in he WW game:
•
A emp ID;
•
A emp s he playe made in a moni o ing sec ion;
•
The ype o pa h (easy o ha d) chosen on each a emp ;
•
Numbe o coins collec ed on each pa h a emp ;
•
Li e los on each pa h a emp ;
•
Time aken o comple e each pa h;
•
T aps isked isi ing and i died o no ;
•
Coins caugh on each isi ed ap.
The TT game collec ed he ollowing da a, o each a emp :

Chap e 9 (Pape 7) - "You Wan o Play a Game?" De ec ing Two Pe sonali y T ai s wi h Sho -Du a ion Mobile Games
250
•
A emp numbe ;
•
ID o he ha d coins caugh ;
•
To al coins caugh ;
•
ID o he diamonds caugh ;
•
Time aken in he a emp ;
•
Final sco e and co esponding ophy.
Using Mic oso Excel 365, he ga he ed a iables we e hen no malized and/o ca ego ized o be
compa able. The ollowing me ics we e calcula ed o he Which Way game:
•
Explo edMS1
(0 - no, 1 - yes) = i he playe wen back in he ini ial scene o explo e moni-
o ing sec ions 1 and/o 2;
•
To alHa dPa hs
(ha d pa hs/ o al ha d pa hs) = a io o ha d pa hs he playe ied o go a
leas once, om a o al o 8 ha d pa hs [0, 1];
•
Insis edHa dPa h
(0, 1 o 2) = 0, i he playe only wen h ough easy pa hs in all bi u ca-
ions; 1, i ied ha d pa h a leas once in only one bi u ca ion, ga e up and wen h ough
he easy pa h he es o he game; 2, i ied ha d pa h one o mo e imes, a leas in wo
bi u ca ions, ga e up and wen h ough he easy pa h, o always ied se e al imes he
ha d pa h un il succeeding;
•
To alT apsRisked
( aps isked/ o al aps) = a io o aps he playe isked going, om a
o al o 5 aps [0, 1];
•
T aps2xA emp s
(0 - no, 1 - yes) = aps he playe ook he chance o go mo e han once,
going o a leas wo di e en aps.
Fo he Time T a el game, he ollowing no malized o ca ego ized a iables we e calcula ed:
•
Ha dCoins
(ha d coins/ o al coins) = a io o ha d coins he playe caugh [0, 1];
•
To alCoins
(caugh coins/ o al coins) = a io o o al coins caugh (ha d + easy) [0, 1];
•
Diamonds
= numbe o diamonds caugh (0, 1, 2, o 3);
•
Time
= ime aken o comple e he game, in seconds;
•
Sco e
=
(𝑑𝑖𝑎𝑚𝑜𝑛𝑑𝑠 × 10+𝑐𝑜𝑖𝑛𝑠) × (10 × 𝑟𝑒𝑤𝑎𝑟𝑑)
, whe e
𝑑𝑖𝑎𝑚𝑜𝑛𝑑𝑠
ep esen he numbe
o diamonds caugh ,
𝑐𝑜𝑖𝑛𝑠
he numbe o coins, and
𝑟𝑒𝑤𝑎𝑟𝑑
he ex a ewa d assigned o
he ime aken o comple e he game (1 i game comple ed in ≤ 400 seconds, 0.9 i 400s
< ime ≤ 450s, 0.8 i 450s < ime ≤ 500s, 0.7 i ime > 500s);
•
T ophies
= 3 is Gold: ≤ 400 seconds o end game & caugh all coins and diamonds. 2 is
Sil e : ≤ 450 seconds & caugh ≥ 2 diamonds. 1 is B onze: ≤ 500 seconds & caugh ≥ 1
diamond. 0 he emaining si ua ions.
Using IBM SPSS S a is ics 26, a co ela ion analysis be ween all pe sonali y ai s and he e e ed
me ics was pe o med on he samples (Pea son o no mal pe sonali y ai s dis ibu ions, Spea -
man o non-no mal). Also, pa ame ic (independen samples - es ) and non-pa ame ic es s
(Mann-Whi ney o 2 g oups and K uskal-Wallis o 3 o mo e g oups), o no mal (o
𝑛
>30) and
Chap e 9 (Pape 7) - "You Wan o Play a Game?" De ec ing Two Pe sonali y T ai s wi h Sho -Du a ion Mobile Games
251
non-no mal dis ibu ions o he pe sonali y ai s, espec i ely, we e pe o med o compa e he
pe sonali y ai s dis ibu ion ac oss ce ain g oups. The esul s and espec i e analysis can be
ound in Sec ion 9.4.
9.4
R
ESULTS AND
A
NALYSIS
9.4.1
Pa icipan s Cha ac e iza ion
Rega ding he ga he ed sample, ou indi iduals had o be emo ed, wo due o da a iles ha we e
no e ie ed om he pa icipan s’ de ice o bo h games, and wo ha did no ill he pe sonali y
ques ionnai e, esul ing in a inal sample o
𝑛
= 100.
As can be obse ed in Table 9.2, he pa icipan s anged om 18 o 58 yea s old. The gende was
equilib a ed, wi h 51% emale and 49% male pa icipan s. 87% o pa icipan s had highe educa-
ion, almos hal (49%) om he Enginee ing & Technology a eas, albei se e al o he a eas we e
also p esen . 61% o pa icipan s we e employed and 48% we e s ill s udying. As o mobile games
playing equency, abou hal he pa icipan s played (53%), 36% used o play bu did no play an-
ymo e, and 11% ne e played, which means he majo i y (89%) had some so o expe ience in
playing. Mobile de ices we e he op played pla o ms (68%), ollowed by lap op/desk op (42%)
almos
ex aequo
wi h gaming consoles (38%). Ad en u e/Ac ion we e he majo i y p e e ed ideo-
game gen es (60%), almos
ex aequo
wi h mul iplaye /RPG/MMORPG (58%), ollowed by s a egy
(53%), and puzzles (44%).
The pa icipan s pe sonali y e ealed he same ype o dis ibu ion as in he p e ious s udies,
which used he BFI (44-i ems) and had mo e pa icipan s (Al es, Ma ins, e al., 2024; Al es,
Ma ins, Sa ai a, e al., 2023) (see Figu e 9.8 o some ele an ai s, and espec i e a iabili y
and loca ion measu es), ha ing esponses abo e he mid-poin o ai s pe cei ably ela ed o sel -
p aising. These ai s e ealed a nega i e skewness, i.e., pa icipan s si ua ed hemsel es mo e
be ween “3- Nei he ag ee no disag ee” and “5-Ag ee s ongly”, e lec ing he same social desi a-
bili y bias common in sel - epo ing ques ionnai es (Ped egon e al., 2012), like o example, he
desi e o being modes (modes y ai ), ambi ious (achie emen -s i ing ai ), cau ious (cau ious-
ness ai ), and adia ing joy (chee ulness ai ).
Chap e 9 (Pape 7) - "You Wan o Play a Game?" De ec ing Two Pe sonali y T ai s wi h Sho -Du a ion Mobile Games
252
Table 9.2. Pa icipan s mos ele an desc ip i e s a is ics (𝑛=100).
n, %
Age
Min = 18, Max = 58, Mean = 32.5 (SD=1.21)
Gende
Female
Male
51
49
Educa ional le el
High school
Bachelo
Deg ee
Mas e
Ph.D.
O he
13
4
47
22
13
1
Fo ma ion a ea
Enginee ing & Technology
Exac Sciences
Medical & Heal h Sciences
Humani ies
Na u al Sciences
Social Sciences
O he
None
49
10
4
11
6
6
13
1
P o essional si ua ion
Employed
S uden
Wo king s uden
Unemployed
O he
49
36
12
2
1
Mobile de ices playing
equency
Se e al imes a day
Once a day
Once a week
Se e al imes a week
Used o play, bu no anymo e
Ne e
Ano he equency
10
7
8
23
36
11
5
Mos equen ly played
pla o ms
Gaming consoles (Xbox, PlayS a ion, e c.)
Lap op/desk op
Po able consoles (Nin endo Swi ch, e c.)
Mobile de ices (sma phone, able , e c.)
Don’ play
Ano he
23
42
15
68
18
3
Some pe cen ages do no ma ch 100% because he ques ion allowed mul iple answe s.
As can also be seen in Figu e 9.8, and con i med by he 1.5xIQR ule – boxplo ule (Vinu ha,
Poo nima, & Saga , 2018), one alue o achie emen -s i ing, and wo o chee ulness and ex-
ci emen seeking a e ou lie s. Howe e , we decided o keep hem, as hese alues ep esen a ue
esponse and e lec impo an indi idual di e ences. Also, we ound ha i s emo al would no
al e he dis ibu ions’ loca ion measu es. Acco ding o he Shapi o-Wilk no mali y es , excep o
exci emen seeking and ad en u ousness, none o he ep esen ed ai s ha e a no mal dis ibu-
ion.
Chap e 9 (Pape 7) - "You Wan o Play a Game?" De ec ing Two Pe sonali y T ai s wi h Sho -Du a ion Mobile Games
253
The a e age pe sonali y sco es we e calcula ed o compa ison, esul ing in con inuous a iables,
o in e p e a ion we conside ed sco es ≤ 2.49 as nega i e (low), be ween 2.50 and 3.49 as neu-
al, and ≥ 3.50 as posi i e (high). The pa icipan s’ ange alls in o a bimodal dis ibu ion, whe e
43% o he pa icipan s conside hey do no ge ang y easily and 30% easily lose hei empe
(median = 2.50). As o anxie y (median = 3.75), 56% a e anxious indi iduals, 38% a e neu al
abou how hey eel, and only 6% do no conside s essing ou easily. Fo he chee ulness ai ,
mos pa icipan s (75%) conside hemsel es o be posi i e indi iduals, lo ing li e, and adia ing joy
(median = 3.75), wi h only 4% hinking he opposi e. The las one clea ly seems o co espond o
he social desi abili y bias associa ed wi h sel - epo ing ques ionnai es. The same can be said o
Figu e 9.8. Pa icipan s’ mean sco es o some pe sonali y ai s in he IPIP-NEO-120 ques ionnai e (𝑛 = 100).
Chap e 9 (Pape 7) - "You Wan o Play a Game?" De ec ing Two Pe sonali y T ai s wi h Sho -Du a ion Mobile Games
254
cau iousness, whe e 72% indi iduals conside hey do no make ash decisions o ac wi hou
hinking (median = 3.88), modes y, whe e 62% conside o be humble pe sons (median = 3.50),
and achie emen -s i ing, wi h 82% conside ing wo king ha d and do mo e han he expec ed (me-
dian = 4.00).
9.4.2
Expe imen s Resul s and Discussion
In he pos -ques ionnai es, 68% and 69% conside ed TT and WW un o play, bu only 9% and 6%
did no ind i un, espec i ely. 44% said TT was “mo e o less easy o play”, and WW 26%. 64% o
pa icipan s said TT made hem lose hei pa ience. 65% and 66% conside ed o eel imme sed,
wi h only 14% and 18% no eeling imme sed, in TT and WW, espec i ely. To some ex en , he
di icul y inhe en o bo h games seemed o hinde imme si eness and un, bu we canno o ge
ha no e e yone likes o play he same so o games, which was no he ocus in his s udy.
Some pa icipan s played wi h low music olume o no olume, which migh also ha e a ec ed
imme si eness. In he WW game, only 12% o he pa icipan s said o go h ough easy pa hs be-
cause hey we e a aid o ying he ha des ones. Mos pa icipan s, 60%, wen h ough he pa h
hey jus el like a he ime.
Some pa icipan s ga e some sugges ions, like he game con ols in WW should be mo e luid.
O he s sugges ed he e should be a sound e ec in TT when hi ing an enemy, and when ca ching
coins, and ha he enemies should be imp o ed o ma ch he scene y s yle. Also, some said TT’s
inal boss should be emo ed, as i implied mo e skill o pass h ough i , leading o se e al playe s
dying a he e y end o he game and s a ing all o e again jus because hey could no unde -
s and how o pass he boss. We ag ee wi h all his, which is also being conside ed in he
minigames cu en ly in de elopmen . Se e al playe s sugges ed he TT game should ha e check-
poin s, so hey didn’ need o s a he game all o e again e e y ime hey los . Howe e , his was
in en ional o es hei pe sis ence. In e es ing sugges ions o pu isible walls in he moni o ing
sec ions o WW, ins ead o he in isible, we e also men ioned, which makes mo e sense and would
ha e imp o ed he gameplay. Al hough manually egis e ed, we also hink he TT minigame should
ha e au oma ically eco ded he me ics e en i he playe did no inish he game, as his could
ha e been an indica o o non-achie e s.

Chap e 9 (Pape 7) - "You Wan o Play a Game?" De ec ing Two Pe sonali y T ai s wi h Sho -Du a ion Mobile Games
255
The pilo s udy helped sol e some bugs and adjus he gameplay, like augmen ing some pla o ms
ha we e oo small o ha d o land on and imp o e he con ols sensi i i y. E en wi h hose im-
p o emen s, se e al pa icipan s in he expe imen s had di icul ies playing, asking o ou help o
o e come ce ain obs acles o e en play he whole game o hem. This esul ed in 13 pa icipan s
(13%) needing us o play o hem in bo h minigames, and 17 in he TT minigame. In an a emp o
a oid bias ela ed o choices, he pa icipan s ga e us ins uc ions on how we should play and wha
i ems o ca ch (“go up”, “jump o he ele a o ”, “ca ch he 3 coins”, e c.), i.e., we only did wha
hey asked us o. This showed he WW was easie o play han he TT game, al hough he WW
game con ols we e no so luid. This was due o he ac he playe s could choose o go only
h ough he easy pa hs, which we e s aigh o wa d, inishing he WW game as e han hose who
ad en u ed h ough ha d pa hs and aps, as he game was no in ended o measu e achie emen -
s i ing. Howe e , consul ing Table 9.3, we can see he e is a clea bias in he imes aken, wi h
he helped pa icipan s gene ally aking less ime o comple e he games. The e o e, and because
we could no de e mine i helped playe s we e somehow in luenced o ac di e en ly due o ou
in e en ion, we decided o exclude hem om he games’ beha io analysis. This esul ed in a
sample o
𝑛
= 86 alid pa icipan s o he WW game beha io analysis
51
, and
𝑛
= 70 o TT.
Table 9.3. Time aken by pa icipan s, in minu es, o comple e he games when helped and no helped.
Game
𝒏
min
max
mean
median
WW
Helped
13
0.88
2.55
1.53
1.34
No helped
86
0.38
4.35
1.88
1.70
TT
Helped
30
1.65
10.9
4.30
3.88
No helped
70
1.63
14.6
4.18
3.90
Fo he TT game, only he ime aken o he i s comple e a emp is conside ed.
9.4.2.1
Cau iousness Acquisi ion
Consul ing Figu e 9.9, Table 9.4, and Appendix E Table E-A.1, Table E-A.2, Table E-A.3, and
Table E-A.6, 10.5% o pa icipan s only wen h ough easy pa hs (
To alHa dPa hs
me ic), 46.5% o
pa icipan s wen a leas once h ough one o wo ha d pa hs, meaning mos pa icipan s chose
easies pa hs. The median was 0.25, he e o e 2 ha d pa hs. As o he
To alT apsRisked
me ic,
36.0% o he pa icipan s did no isk going o a ap, 47.7% wen only o one ap, 14.0% o only 2,
and 2.3% o 3 aps. Rega ding
Explo edMS1
me ic, 64% o pa icipan s did no explo e moni o ing
51
The CSV ile was no e ie ed o a pa icipan in he WW game.
Chap e 9 (Pape 7) - "You Wan o Play a Game?" De ec ing Two Pe sonali y T ai s wi h Sho -Du a ion Mobile Games
256
sec ions 1 and/o 2, and 36% did. The ime aken o comple e he game a ied om 23s o 261s,
wi h a median o 102s.
Nex a e p esen ed he co ela ions and independen samples es s esul s o he WW game and
p oposed hypo heses.
H1 - A mo e cau ious pe son will a oid ha d pa hs, despi e knowing she can ge mo e
ewa ds
No s a is ically signi ican co ela ions we e ound be ween cau iousness and he numbe o ha d
pa hs aken. Howe e , s a is ically signi ican co ela ions be ween
To alHa dPa hs
and ange ,
modes y, and chee ulness, we e ound (Table 9.4). This shows pa icipan s who seek ha de
pa hs a e indi iduals who ge ang y less easily and a e he e o e mo e pa ien , a e less modes
and end o conside hemsel es be e han o he s, and like o ha e un.
To de e mine i he pa icipan s’ cau iousness was di e en be ween hose who p e e ed mos ly
easy pa hs (a mos 2 ha d pa hs), and hose who ied mo e ha d pa hs, he a iable was spli in o
wo g oups, whe e
𝐺
0
< 3 ha d pa hs and
𝐺
1
≥ 3 ha d pa hs, and an independen -samples Mann-
Whi ney’s es was pe o med (Figu e 9.10a). The means o cau iousness we e no s a is ically
Figu e 9.9. His og ams o he me ics measu ed in he Which Way minigame
(𝑛=86).
Chap e 9 (Pape 7) - "You Wan o Play a Game?" De ec ing Two Pe sonali y T ai s wi h Sho -Du a ion Mobile Games
257
di e en ac oss he wo g oups (
𝑝
= 0.709), no ac oss all he eigh possible
To alHa dPa h
alues
(
𝑝
= 0.736) (Figu e 9.10b). The e o e, hypo hesis H1 could no be suppo ed.
Table 9.4. Co ela ion coe icien s be ween he WW game measu ed a iables and ele an pe sonali y ai s (𝑛 = 86).
T ai
No mali y
(Shapi o-Wilk)
Time
To alHa d
Pa hs
Insis ed
Ha d
Pa h
To alT aps
Risked
T aps2xA emp s
Explo ed
MS1
O3-Emo ionali y
No
0.025
-0.033
-0.049
-0.042
-0.071
-0.189
C6-Cau iousness
No
0.129
-0.018
-0.050
-0.060
-0.145
0.046
E6-Chee ulness
No
0.176
0.219*
0.242*
0.047
-0.099
0.154
A1-T us
No
-0.100
0.130
0.193
0.036
0.014
-0.007
A5-Modes y
Yes
-0.076
-0.211*
-0.195
-0.137
-0.033
-0.162
N1-Anxie y
No
-0.222*
-0.057
-0.043
-0.070
0.005
-0.123
N2-Ange
No
-0.197
-0.224*
-0.109
-0.095
-0.050
-0.097
N6-Vulne abili y
No
-0.178
-0.155
-0.092
-0.207
0.021
-0.025
S a is ically signi ican alues a e in bold [*Co ela ion is signi ican a he 0.05 le el (2- ailed)]. Unde lined alues ha e a p- alue close o 0.05. A Pea -
son co ela ion was calcula ed o he pe sonali y no mal dis ibu ions (in i alics). A Spea man co ela ion was calcula ed o non-no mal dis ibu ions.
H2 – A mo e cau ious pe son will a oid en e ing he aps, e en knowing he can ge
mo e coins
No s a is ically signi ican co ela ions we e ound be ween cau iousness and en e ing a ap (
To-
alT apsRisked
), no di e en cau iousness’ means be ween he pa icipan s who did no en e a
ap,
𝐺
0
, and he ones who en e ed a leas one ap,
𝐺
1
(Appendix E Table E-A.7). The e o e, hy-
po hesis H2 could no be suppo ed.
Rega ding o he pe sonali y ai s, di e en means ac oss he e e ed wo g oups we e ound. The
pa icipan s who a oided en e ing he aps had an a e age lowe sel -discipline, ha ing mo e di i-
cul y in s a ing asks; we e mo e ulne able, panicking mo e easily; less asse i e, wai ing o o h-
e s o ake he lead; and mo e modes .
As only 3 pa icipan s isked going o a leas wo di e en aps mo e han once (
T aps2xA emp s
,
Appendix E Table E-A.4), he me ic did no ha e su icien da a o be ela ed o he cau iousness
ai o o he pe sonali y ai s.
H3 – I a mo e cau ious pe son en e s a ha d pa h and dies, a e espawning, she
will choose he easy pa h ins ead
Rega ding he pe sis ence o y a ha d pa h (
Insis edHa dPa h
, Figu e 9.9 and Appendix E Table
E-A.5), 10.5% only wen h ough easy pa hs in all bi u ca ions (
𝐺
0
), 19.8% ied he ha d pa h a
leas once in only one bi u ca ion, ga e up and wen h ough he easy pa h he es o he game
(
𝐺
1
), 69.8% ied he ha d pa h one o mo e imes, a leas in wo bi u ca ions, ga e up and wen
Chap e 9 (Pape 7) - "You Wan o Play a Game?" De ec ing Two Pe sonali y T ai s wi h Sho -Du a ion Mobile Games
258
Figu e 9.10. Independen -samples es s o he cau iousness dis ibu ion ac oss di e en g oups o he WW
me ics (𝑛=86).
h ough he easy pa h, o always ied se e al imes he ha d pa h un il succeeding (
𝐺
2
). A e he
pa icipan s who ga e up a e dying in a ha d pa h (
𝐺
1
) mo e cau ious han hose who insis ed
mo e on ha d pa hs (
𝐺
2
)? Compa ing he wo g oups (Figu e 9.10c), he means o cau iousness
a e he same ac oss he wo g oups (
𝑝
= 0.792), he e o e, he pa icipan s who ga e up ha d
pa hs, going h ough easy pa hs ins ead, a e no mo e cau ious han hose who insis ed on ha d
pa hs. And
𝐺
1
’s pa icipan s a e mo e cau ious han
𝐺
0
’s? No, as can be seen in he same g aph-
ic. The same esul happens i we compa e he means ac oss all he eigh possible
To alHa dPa h
alues (
𝑝
= 0.736) (Figu e 9.10b), and i we join
𝐺
0
wi h
𝐺
1
( enamed
𝐺
0
o simplici y) and
compa e wi h
𝐺
2
(Figu e 9.10d). Fu he mo e, no s a is ically signi ican co ela ions we e ound
be ween cau iousness and he pe sis ence o ake ha d pa hs. The e o e, hypo hesis H3 could no
be suppo ed.
(a) (b) (c)
(d) (e) ( )
Chap e 9 (Pape 7) - "You Wan o Play a Game?" De ec ing Two Pe sonali y T ai s wi h Sho -Du a ion Mobile Games
265
A cu ious posi i e ela ionship as well was obse ed be ween iendliness and he diamonds
caugh , meaning indi iduals who gain mo e ophies hough o hemsel es as being able o make
iends easily and eel com o able a ound o he s. Possibly, he pa icipan s wi h high sel - epo ed
ex a e sion ( iendliness) in en ionally wan ed mo e o engage in his s udy, hus showing mo e
disposi ion o accomplish he gi en asks and he e o e mee he game objec i es, and conse-
quen ly he esea che s’ expec a ions, which ul ima ely is a way o seeking social in e ac ion. In
his ega d, ex a e sion wo ks in a way ha would mo e o en be a ibu able o Ag eeableness,
bu ha in his speci ic gaming con ex , bo h dimensions disambigua ion canno be assu ed.
An (ob ious) ela ionship wi h exci emen seeking and ad en u ousness was encoun e ed, being
posi i ely co ela ed wi h all conside ed me ics (excep ime o exci emen ). Indi iduals who like o
eel exci emen and ad en u e a e he ones who s i e mo e o ge all he collec ables and be e
sco es in he game. This is in line wi h Teng and Chen (2009) esul s, who ound exci emen seek-
ing and achie emen we e posi i ely ela ed o challenge. Gimmon, Rob, Fa ja, and Golan (2023)
disco e ed ha expe ienced escape oom playe s played o he exci emen and he less expe i-
enced o achie emen . Being chee ul and posi i e is posi i ely ela ed o ge ing mo e coins, dia-
monds, highe sco es, and be e ophies.
No signi ican co ela ion be ween exci emen seeking and achie emen -s i ing was ound, mean-
ing hey migh no be co ela ed as ini ially expec ed. I we look closely a he ai s’ de ini ion,
someone who is pe ec ionis and wo ks ha d is no necessa ily he same as a pe son who seeks
ad en u e. In ac , exci emen seeking is mo e connec ed o less conscien ious indi iduals and
achie emen o mo e conscien ious ones (Al es, Ma ins, Sa ai a, e al., 2023).
And ad en u e? Is i posi i ely ela ed o achie emen -s i ing? Cos a J e al. (1995) conside
someone posi i e in he openness o ac ions ai (O4) is an indi idual who needs a ie y and ge s
bo ed by ou ine, co esponding o he ad en u ousness ai conside ed by Johnson (2014) in he
IPIP-NEO-120 scale. Someone posi i e in achie emen -s i ing is an ambi ious indi idual who en-
dea o s o excellence, ha ing high s anda ds, acco ding o Cos a J e al. (1995), and wo ks ha d
doing mo e han he expec ed (Johnson, 2014). A so o associa ion wi h bo h ai s seems e i-
den , which is suppo ed by hei posi i e co ela ion.

Chap e 9 (Pape 7) - "You Wan o Play a Game?" De ec ing Two Pe sonali y T ai s wi h Sho -Du a ion Mobile Games
266
All hese esul s show he TT game can be used o de e mine achie emen -s i ing and o he pe -
sonali y ai s. This is a g ea ad ance o s a imp o ing and e ining pe sonali y games o implici -
ly de ec pe sonali y ai s, which we a e al eady wo king on
52
.
9.5
C
ONCLUSIONS AND
F
UTURE
W
ORK
Pe sonali y is a s ong in luence on he use s’ choices in a a ie y o domains, such as ou ism,
music, mo ies, e c. Howe e , o ocus only on he i e dimensions o pe sonali y, and no on he
hi y ai s and on he co ela ions be ween hem, may limi he p e e ences p edic ion. Fo exam-
ple, a pe son conside ed ex a e ed may no be a isk ake o like ad enaline ac i i ies. I would
no be e y good i a RS sugges ed a olle coas e o he ou is . Also, o acqui e someone’s pe -
sonali y is complex and a challenge, as he explici echniques, such as sel - epo ing ques ion-
nai es, a e subjec o he social desi abili y bias and ake esponses, hey a e bo ing and ime-
consuming o ill, and subjec o w ong in e p e a ions; he implici ones need a g ea amoun o
use s’ in e ac ions o make he p edic ions, use complex and/o no so accu a e games, a e mos -
ly only o desk op compu e s, and/o only p edic he b oade i e pe sonali y dimensions. Also, o
he bes o ou knowledge, he e a e no sho -du a ion mobile games ha can accu a ely p edic
pe sonali y.
To y o o e come hose limi a ions, we de eloped wo mobile minigames, “Time T a el Mania”
and “Which Way”, as concep p oo o implici ly de ec ing wo o he Big Fi e pe sonali y mo e
g anula ai s, achie emen -s i ing and cau iousness, espec i ely, in a sho playing ime (< 5
min) and a a i s in e ac ion. Con a y o F. Y. Wu e al. (2022) s udy, TT game p o ed o be ca-
pable o measu ing achie emen -s i ing, easily iden i ying he achie e s and non-achie e s as
sugges ed by Ba le (1996), suppo ing hypo heses H6 o H9, showing wi h he igh kind o games
i may be possible o accu a ely measu e hose ai s. Howe e , he WW game could no success-
ully measu e cau iousness, no suppo ing hypo heses H1 o H5, sugges ing ha mo e pa ici-
pan s we e needed o ob ain signi ican co ela ions (a leas la ge on he lowe sco es o he ai ),
which should be conside ed in u u e s udies.
52
O he esul s could be analyzed in he ob ained da a, bu only he mos ele an we e p esen ed.
Chap e 9 (Pape 7) - "You Wan o Play a Game?" De ec ing Two Pe sonali y T ai s wi h Sho -Du a ion Mobile Games
267
Thus, we hink he main limi a ion o his s udy was no he games concep hemsel es, bu he
sample size o ga he g ound u h da a o he Openness o Expe ience, Conscien iousness, and
Ag eeableness pe sonali y dimensions, as also e i ied, o example, by T. T. Nguyen e al. (2018)
and Al es, Ma ins, Sa ai a, e al. (2023), and consequen ly o some o hei co esponding ai s,
such as he ones s udied. These a e de ini ely he ha des dimensions o measu e, and we a e
cu en ly wo king on ways o ying o o e come ha .
An unexpec ed inding was ha TT was measu ing cau iousness, which was ela ed o he playe s
who wen o long and deadly pa hs o collec ha d coins and diamonds. So, cau iousness is mo e
ela ed o being p uden and hough ul (Cos a J e al., 1995) han o being ea ul, which migh
be di icul o measu e sepa a ely in a game. This di e ence needs o be u he s udied in he
u u e, so cau iousness can be accu a ely measu ed.
We also ound some pa icipan s had di icul y playing bo h games, o jus TT game. Al hough de-
signed o imply less skill, he games needed o be e en easie o play, so any ype o pe son could
play i , as well as mo e imme si e (such as ha ing sound e ec s in ce ain e en s, as sugges ed by
some pa icipan s, and o cing pa icipan s o hea he game music wi h headphones), o os e
pa icipan s’ enjoymen and in ol emen . This could also ha e been a limi ing ac o o he esul s
ob ained, especially o he cau iousness game, whe e a mo e imme si e en i onmen could os e
mo e p udence and ale ness in he playe s. We also hink he bes accomplishmen o each col-
lec ible in TT game should ha e been eco ded, and ha he pa icipan s shouldn’ ha e been
obliged o inish he game, as i could ha e e ealed s onge co ela ions o achie emen -s i ing.
In e es ingly, we also ound s a is ically signi ican co ela ions o o he ai s. Pa icipan s who ge
ang y less easily and a e less modes chose he mos challenging pa hs in he WW game. Also, in
he TT game, he pa icipan s who ge less ang y a e he ones who go mo e collec ables and be -
e sco es. The same can be said o exci emen seeking, chee ulness, and ad en u ousness.
These indings make sense and a e in line wi h he esul s ob ained by Teng and Chen (2009) o
exci emen seeking and achie emen , and wi h he obse a ions made a he end o Sec ion 9.2.2,
as se e al pe sonali y ai s can be ela ed o each o he and may no be possible o be measu ed
sepa a ely. These indings a e an in e es ing subjec o u u e s udies in he a ea.
Chap e 9 (Pape 7) - "You Wan o Play a Game?" De ec ing Two Pe sonali y T ai s wi h Sho -Du a ion Mobile Games
268
Wi h his wo k, o he bes o ou knowledge, we could measu e mo e g anula pe sonali y ai s
and wi h a much smalle du a ion play ime and in e ac ions han he games ound in li e a u e. We
belie e his is a g ea ad ance in he a ea, because, beside emo ing he social desi abili y bias in
a ew minu es wi h simple minigames, he use s will be mo e engaged and amused while hei
de ailed pe sonali y ai s a e acqui ed, which is no he case o illing explici ques ionnai es.
We can he e o e a i m we a e on he igh pa h o de elop sho -du a ion se ious mobile games o
implici ly acqui e an indi idual’s mo e g anula pe sonali y ai s, wi hou he need o ML ech-
niques. Wi h he igh design and me ics, his kind o games can be in eg a ed in o pe sonali y-
based sys ems o help ga he he use s’ pe sonali y and sol e he limi a ions ound in he a ea, as
well as be used in o he a eas, such as o assess and selec job in e iewees o example.
Funding
This wo k was suppo ed by he PRR – ATT P ojec unde G an C645192610-00000060; PRR
and Nex Gene a ion EU Eu opean Funds; and Po uguese Founda ion o Science and Technology
(FCT) unde G an s UIDB/00319/2020 and UIDB/00760/2020. The wo k o Pa ícia Al es was
suppo ed by he Eu opean Social Fund and FCT unde G an 2020.06129.BD.
Acknowledgmen s
The au ho s would like o hank Supe Bock G oup, S.A., o sponso ing he expe imen s; Ag upa-
men o de Escolas Gaia Nascen e, Gaiu b, and T us - Ges ão In eg ada de Saúde, S.A. o hei
help in ga he ing mo e pa icipan s and allowing o conduc he expe imen s locally wi h hei col-
labo a o s; and Ped o Oli ei a om ISLab o his a ailabili y, sympa hy, and suppo in dissemina -
ing and helping conduc he s udy a Uni e si y o Minho.
E hics s a emen
This wo k in ol ed human subjec s in i s esea ch. App o al o all e hical and expe imen al p oce-
du es and p o ocols was g an ed by he Poly echnic o Po o Da a P o ec ion O ice, and pe o med
in line wi h he ins i u ion equi emen s. The used ques ionnai es we e anonymous and con iden-
ial, and he collec ed da a was analyzed in agg ega ed o m. All pa icipan s signed an in o med
Chap e 9 (Pape 7) - "You Wan o Play a Game?" De ec ing Two Pe sonali y T ai s wi h Sho -Du a ion Mobile Games
269
consen o pa icipa e in he s udy, ensu ing con iden iali y, anonymi y, and in o ming he aims o
he esea ch and whe e he da a was s o ed.
Supplemen al online ma e ial
The ques ionnai es used in he expe imen s can be consul ed in he olde “Ques ionnai es” o he
supplemen a y ma e ial (in Po uguese). The comple e collec ed esponses and espec i e s a is-
ics, as well as he collec ed games’ me ics (da ase used), can be ound in olde “Responses” o
he supplemen a y ma e ial.
Supplemen a y games’ s a is ics and he comple e Pea son and Spea man’s co ela ions ables
can be ound in he supplemen a y ma e ial main olde .
The games APK can be downloaded in he supplemen a y ma e ial (APK olde ). Please accep he
wa nings in o de o ins all.
Pa III
IN CONCLUSION

271
“Logic will ge you om A o B. Imagina ion will ake you e e ywhe e.”
Albe Eins ein
272
10
C O N C L U S I O N S
This doc o al wo k p oposed o add ess se e al cu en main challenges o G oup Recommende
Sys ems, some pe aining o he ou ism a ea, namely:
1.
The cold-s a p oblem;
2.
The g oup’s membe s’ he e ogeneous and con lic ing p e e ences;
3.
Show conce n o he indi idual use s’ needs;
4.
The in usi eness o build he use s’ ini ial p o ile;
5.
Spa si y issues;
6.
How he use s’ p e e ences can be agg ega ed so all membe s can be sa is ied wi h he
ecommenda ions;
7.
How he g oup can be aided in eaching he inal choice wi hou he emo ional con agion
phenomenon;
8.
How o p omo e he sense o communi y and companionship in excu sion g oups.
Based on he e idence ound in li e a u e, all hese challenges led o many indi idual and g oup
b ains o ming sessions on how such p oblems could be add essed. Se e al e sions o a GRS p o-
o ype we e idealized un il he inal e sion was inally concep ualized.
In a i s concep , he whole p ocess om he planning o he excu sion i sel was being consid-
e ed, including he use o an AR a a a o ep esen he ou is , and dialogue games be ween
agen s o acili a e he ecommenda ions nego ia ion p ocess, which is e lec ed in Pape 1 (Chap-
e 3). Bu he ideas we e s ill esh, and due o he de ined imeline, i was ecognized ha no
e e y hing could be conside ed in he doc o al wo k. So, as he esea ch p og essed, he engi-
nee ed ideas shi ed o a mo e cohesi e and highe quali y solu ion di ec ed o sol e he iden i ied
p oblems mos ly in he phase o he excu sion planning.
Chap e 10 – Conclusions
273
The impo ance o pe sonali y in p edic ing he use s’ p e e ences in a a ie y o domains was
acknowledged and i could be he missing le e age in he p ojec , as was seen in Pape s 2 and 3
(Chap e 4 and Chap e 5, espec i ely). A la ge-scale s udy wi h eal pa icipan s was conduc ed,
showing i was possible o p edic ou is p e e ences, mo i a ions and a el- ela ed p e e ences &
conce ns based on he aw pe sonali y dimensions o a pe son ins ead o using ou is ypologies
like mos o he exis ing wo ks, which answe ed o esea ch ques ions “RQ1.1 - Does pe sonali y
p edic p e e ences o ou is a ac ions, a el mo i a ions, and pe sonal p e e ences & conce ns
when a elling?” and “RQ1.2 - I so, which pe sonali y dimensions o ai s a e an eceden s o
which ou is a ac ions, a el mo i a ions, and pe sonal p e e ences & conce ns when a el-
ling?”. This esul ed in h ee models ha we e used as he basis o p edic ing he ou is s’ p e e -
ences in he GRS p o o ype, being he i s , o he bes o ou knowledge, o p opose such models,
sol ing he cold-s a p oblem, as he e was no need o da a mine a g ea numbe o use s’ da a o
in e ac ions o make he ini ial ecommenda ions. The de eloped pe sonali y models allowed o
p edic he ou is s’ p e e ences in ele en ou ism ca ego ies, i e a elling mo i a ions, and ou
a el- ela ed p e e ences & conce ns, only by knowing hei aw pe sonali y, using he mos used
and ecognized pe sonali y model, he Big Fi e.
The i s wo king GRS p o o ype was demons a ed in Pape 4 (Chap e 6). The use o a mic o-
se ices-based a chi ec u e, including a Mul i-Agen Mic ose ice wi h agen s modeled wi h he
ou is s’ cha ac e is ics, we e he co e ea u es o he demons a ion. This a chi ec u e was
hough up o make he GRS modula , independen , scalable, easily main ainable, so i could be
used by any HTTP clien , bene i ing om he agen s being exposed as esou ces ha could be
easily accessed, and a mo e e icien in e ope abili y, p o iding as e ecommenda ions. By using
he de eloped pe sonali y models, pe sonalized and sa is ac o y ecommenda ions could be gen-
e a ed, no jus o excu sion g oups, bu also o indi iduals, showing conce n o he indi idual
ou is s needs (such as hei physical limi a ions, ea s/phobias, and a el- ela ed p e e ences &
conce ns), as demons a ed in a eal use case scena io (Pape 5, Chap e 7). The ini ial ecom-
menda ions a e solely based on he use s’ pe sonali y, wi hou needing o da a mine la ge
amoun s o da a and use s’ in e ac ions, also sol ing he cold-s a p oblem. All his answe s o
esea ch ques ions “RQ2 – Is i possible o gene a e pe sonalized and ele an ecommenda ions
Chap e 10 – Conclusions
274
based on he ou is s’ pe sonali y?” and “RQ3 - Can he cold-s a p oblem be sol ed jus by using
he ou is s’ pe sonali y?”.
To sol e he he e ogeini y and con lic ing p e e ences in excu sion g oups, especially in la ge
ones, he idea o di iding he main g oup in o subg oups eme ged. As i was ound ha pe sonali y
is s ongly ela ed o ou is p e e ences, i simila pe sonali y ou is s could be g ouped oge he ,
ins ead o being mixed in he main g oup, he he e ogeini y and con lic ing p e e ences p oblem
could be ackled. So, he idea o c ea ing clus e s o simila pe sonali y ou is s by he MAMS
eme ged (demog aphic il e ing). These clus e s needed o ha e a es ic ion, o he p oblem would
con inue. So, i was decided o include only ou is s wi h 0.80
∈
[0,1] o mo e simila i y in a clus-
e , which was shown o be he ideal alue in he expe imen s pe o med (see Pape 5, Chap e 7).
Also, o be e and as e assign he ou is s in o a clus e , ha p ocess is au oma ically pe o med
a he ou is ’s egis a ion in he applica ion, meaning when a ou is egis e s in he GRS, i a
simila i y o he cen oid o 0.80 o mo e is ound, she is immedia ely assigned o a clus e , o he -
wise, she gene a es a new clus e , becoming he cen oid. The clus e s a e c ea ed in eal- ime as
he ou is s egis e . This led o a new dynamic clus e ing algo i hm,
𝑑
-means, based in a modi ied
e sion o
𝑘
--means, ha does no need o know and ini ialize he numbe o ini ial clus e s
a p io i
,
and, o he bes o ou knowledge, is a no el y in GRS. The implemen ed clus e ing algo i hm ou -
pe o med
𝑘
--means and
𝑘
-means++ in scalabili y and clus e ing quali y, and signi ican ly su -
passed
𝑘
-means++ in he ime needed o c ea e clus e s in bulk.
𝑑
-means a oids was ing ime on
clus e ini ializa ion o pe o ming clus e ing du ing o line and online phases. Ins ead, clus e s a e
c ea ed dynamically when needed, o he new use is di ec ly added o an exis ing clus e , wi h i s
cen oid being upda ed acco dingly.
𝑑
-means ensu es ha he e a e no ou lie s o la ge dis ances
be ween clus e membe s by assigning ou is s o an exis ing clus e only i hei simila i y o he
clus e 's cen oid is a leas 0.80 (con igu able). This app oach makes he clus e ing p ocess mo e
p ecise and as e , as i is un in he MAMS. Tou is s who do no mee his c i e ion a e placed in a
new clus e , which can la e accommoda e o he simila ou is s as mo e use s egis e in he app,
helping o mi iga e spa si y issues and ackling he dynamic shi s in he g oup’s composi ion wi h-
ou a ec ing he ecommenda ions quali y.
Re e ences
281
In e na ional Con e ence, PAAMS 2022, L'Aquila, I aly, July 13–15, 2022,
P oceedings.
Al es, P., Ma ins, A., Neg ão, F., No ais, P., Almeida, A., & Ma ei os, G. (2024). A e
He e ogeini y and Con lic ing P e e ences No Longe a P oblem? Pe sonali y-Based
Dynamic Clus e ing o G oup Recommende Sys ems. Expe sys ems wi h
applica ions, 255, Pa D. doi:h ps://doi.o g/10.1016/j.eswa.2024.124812
Al es, P., Ma ins, A., No ais, P., & Ma ei os, G. (2023, 2023). Imp o ing G oup
Recommenda ions using Pe sonali y, Dynamic Clus e ing and Mul i-Agen
Mic oSe ices. Pape p esen ed a he P oceedings o he 17 h ACM Con e ence on
Recommende Sys ems.
Al es, P., Ma ins, H., Sa ai a, P., Ca nei o, J., No ais, P., & Ma ei os, G. (2023). G oup
Recommende Sys ems o Tou ism: How does Pe sonali y p edic P e e ences o
A ac ions, T a el Mo i a ions, P e e ences and Conce ns? Use Modeling and Use -
Adap ed In e ac ion. doi:h ps://doi.o g/10.1007/s11257-023-09361-2
Al es, P., Sa ai a, P., Ca nei o, J., Campos, P., Ma ins, H., No ais, P., & Ma ei os, G. (2020).
Modeling Tou is s' Pe sonali y in Recommende Sys ems: How Does Pe sonali y
In luence P e e ences o Tou is A ac ions? Pape p esen ed a he P oceedings o
he 28 h ACM Con e ence on Use Modeling, Adap a ion and Pe sonaliza ion.
Al es, P., T indade, J., Mon ei o, G., Sa ai a, P., Campos, P., Ma ei os, G., & No ais, P. (2024).
"You Wan o Play a Game?" De ec ing Pe sonali y T ai s wi h Mobile Minigames.
TechRxi . Re ie ed om 10.36227/ ech xi .171177596.69544930/ 1
Ama o, A. (2023). Recommende Sys ems in he Museum Sec o : An O e iew. Pape p esen ed
a he Ad anced In o ma ion Ne wo king and Applica ions: P oceedings o he 37 h
In e na ional Con e ence on Ad anced In o ma ion Ne wo king and Applica ions
(AINA-2023), Volume 3.
Anadón, X., Sanahuja, P., T a e , V. J., Lopez, A., & Ribelles, J. (2021). Cha ac e ising playe s o
a cube puzzle game wi h a wo-le el bag o wo ds. Pape p esen ed a he Adjunc
P oceedings o he 29 h ACM Con e ence on Use Modeling, Adap a ion and
Pe sonaliza ion.
Ande son, A. H., Bade , M., Ba d, E. G., Boyle, E., Dohe y, G., Ga od, S., . . . Mille , J. (1991).
The HCRC map ask co pus. Language and speech, 34(4), 351-366.
A dissono, L., Goy, A., Pe one, G., Segnan, M., & To asso, P. (2003). In igue: pe sonalized
ecommenda ion o ou is a ac ions o desk op and hand held de ices. Applied
a i icial in elligence, 17(8-9), 687-714.
A i , Y. M., Ayunda, N., Diah, N. M., & Ga cia, M. B. (2024). A Sys ema ic Re iew o Se ious
Games o Heal h Educa ion: Technology, Challenges, and Fu u e Di ec ions.
T ans o ma i e App oaches o Pa ien Li e acy and Heal hca e Inno a ion, 20-45.
A hu , D., & Vassil i skii, S. (2007). k-means++: The ad an ages o ca e ul seeding. Pape
p esen ed a he Soda.
Aschwanden, D., S ickhouse , J. E., Seske , A. A., Lee, J. H., Luche i, M., S ephan, Y., . . .
Te acciano, A. (2021). Psychological and beha iou al esponses o co ona i us
disease 2019: The ole o pe sonali y. Eu opean jou nal o pe sonali y, 35(1), 51-66.
Ayala-Gómez, F., Da óczy, B. Z., Ma hioudakis, M., Benczú , A., & Gionis, A. (2017). Whe e
could we go? Recommenda ions o g oups in loca ion-based social ne wo ks.
Ayemowa, M. O., Ib ahim, R., & Khan, M. M. (2024). Analysis o Recommende Sys em Using
Gene a i e A i icial In elligence: A Sys ema ic Li e a u e Re iew. IEEE Access.
Bansal, S., & Sha ma, R. (2024). Re ealing he E olu ion o Ne lix Recommende Sys ems.
Pape p esen ed a he 2024 11 h In e na ional Con e ence on Compu ing o
Sus ainable Global De elopmen (INDIACom).

Re e ences
282
Ba -El, D., & E. Ringland, K. (2020). C a ing game-based lea ning: An analysis o lessons o
minec a educa ion edi ion. Pape p esen ed a he P oceedings o he 15 h
In e na ional Con e ence on he Founda ions o Digi al Games.
Ba kan, O., Bogina, V., Gu e i ch, L., Ashe , Y., & Koenigs ein, N. (2024). A Coun e ac ual
F amewo k o Lea ning and E alua ing Explana ions o Recommende Sys ems.
Pape p esen ed a he P oceedings o he ACM on Web Con e ence 2024.
Ba na, B., & Fodo , S. (2018). Gami ica ion’s impac on employee engagemen : Enhancing
employee well-being wi h a cloud based gami ied eam-building applica ion. Pape
p esen ed a he 2018 6 h In e na ional Con e ence on Fu u e In e ne o Things and
Cloud Wo kshops (FiCloudW).
Ba le, R. (1996). Hea s, clubs, diamonds, spades: Playe s who sui MUDs. Jou nal o MUD
esea ch, 1(1), 19.
Basala, S. L. (1997). In es iga ing a el s yle p e e ences o isi ing a no el des ina ion: A
conjoin -based app oach. Pu due Uni e si y,
Basala, S. L., & Klenosky, D. B. (2001). T a el-s yle p e e ences o isi ing a no el
des ina ion: A conjoin in es iga ion ac oss he no el y- amilia i y con inuum.
Jou nal o T a el Resea ch, 40(2), 172-182.
Ba e , M., Mo eno, A., Sánchez, D., Ise n, D., & Valls, A. (2012). Tu is @: Agen -based
pe sonalised ecommenda ion o ou is ac i i ies. Expe sys ems wi h applica ions,
39(8), 7319-7329.
Ba inca, L., Lep i, B., Mana, N., & Pianesi, F. (2012). Mul imodal ecogni ion o pe sonali y
ai s in human-compu e collabo a i e asks. Pape p esen ed a he P oceedings o
he 14 h ACM in e na ional con e ence on Mul imodal in e ac ion.
Baue , R. A. (1960). Consume beha io as isk aking. Pape p esen ed a he P oceedings o
he 43 d Na ional Con e ence o he Ame ican Ma ke ing Assoca ion, June 15, 16, 17,
Chicago, Illinois, 1960.
Becken, S. (2010). The impo ance o clima e and wea he o ou ism: li e a u e e iew.
Bekke man, P., K aus, S., & Ricci, F. (2006). Applying coope a i e nego ia ion me hodology o
g oup ecommenda ion p oblem. Pape p esen ed a he P oceedings o Wo kshop on
Recommende Sys ems in 17 h Eu opean Con e ence on A i icial In elligence (ECAI
2006).
Bellman, R. (1957). Dynamic P og amming: P ince on Uni e si y P ess.
Bengali, V., & Veld , N. (2023). Fas e App oxima ion Algo i hms o Pa ame e ized G aph
Clus e ing and Edge Labeling. Pape p esen ed a he ACM In e na ional Con e ence
on In o ma ion and Knowledge Managemen (CIKM ’23), Bi mingham, Uni ed
Kingdom.
Benne , J., & Lanning, S. (2007). The ne lix p ize. Pape p esen ed a he P oceedings o KDD
cup and wo kshop.
Be dun, F. D., & A men ano, M. G. (2018). Modeling Use s' Collabo a i e Beha io Wi h a
Se ious Game. IEEE T ansac ions on Games, 11(2), 121-128.
Bha ia, S., & Ryan, A. M. (2018). Hi ing o he win: Game-based assessmen in employee
selec ion. J. H. Dulebohn & D. L. S one, The b a e new wo ld o eHRM 2.0, 81-110.
Bi, X., Qu, A., Wang, J., & Shen, X. (2017). A g oup-speci ic ecommende sys em. Jou nal o he
Ame ican S a is ical Associa ion, 112(519), 1344-1353.
Bi k, M. V., Toke , D., Mand yk, R. L., & Cona i, C. (2015). Modeling mo i a ion in a social
ne wo k game using playe -cen ic ai s and pe sonali y ai s. Pape p esen ed a he
Use Modeling, Adap a ion and Pe sonaliza ion: 23 d In e na ional Con e ence,
UMAP 2015, Dublin, I eland, June 29--July 3, 2015. P oceedings 23.
Blanco, J. M., Ge, M., & Pi ne , T. (2023). An Adap i e Fil e o P e e ence Fine-Tuning in
Recommende Sys ems. Pape p esen ed a he Web In o ma ion Sys ems and
Re e ences
283
Technologies: 16 h In e na ional Con e ence, WEBIST 2020, No embe 3–5, 2020,
and 17 h In e na ional Con e ence, WEBIST 2021, Oc obe 26–28, 2021, Vi ual
E en s, Re ised Selec ed Pape s.
Bobadilla, J., O ega, F., He nando, A., & Be nal, J. (2012). A collabo a i e il e ing app oach o
mi iga e he new use cold s a p oblem. Knowledge-Based Sys ems, 26, 225-238.
Boksbe ge , P. E., & Laesse , C. (2008). Segmen ing he senio a el ma ke by means o
a el mo i a ion-Insigh s om a ma u e ma ke (Swi ze land).
Bo a o, L., & Ca a, S. (2010). S a e-o - he-a in g oup ecommenda ion and new
app oaches o au oma ic iden i ica ion o g oups. In In o ma ion e ie al and
mining in dis ibu ed en i onmen s (pp. 1-20): Sp inge .
Bo a o, L., & Ca a, S. (2015). ART: g oup ecommenda ion app oaches o au oma ically
de ec ed g oups. In e na ional Jou nal o Machine Lea ning and Cybe ne ics, 6(6),
953-980.
Bo kowski, P., Jażdżewska-Gu a, M., & Szmel e -Ja osz, A. (2021). Lockdowned: E e yday
mobili y changes in esponse o COVID-19. Jou nal o T anspo Geog aphy, 90,
102906.
Bo às, J., Mo eno, A., & Valls, A. (2014). In elligen ou ism ecommende sys ems: A su ey.
Expe sys ems wi h applica ions, 41(16), 7370-7389.
B eese, J. S., Hecke man, D., & Kadie, C. (2013). Empi ical analysis o p edic i e algo i hms
o collabo a i e il e ing. a Xi p ep in a Xi :1301.7363.
B yman, A., & C ame , D. (1992). Quan i a i e da a analysis o social scien is s. Es udios
Geog á icos, 53(207), 347.
Bujisic, M., Bilgihan, A., & Smi h, S. (2015). Rela ionship be ween gues expe ience,
pe sonali y cha ac e is ics, and sa is ac ion: Mode a ing e ec o ex a e sion and
openness o expe ience. Tou ism Analysis, 20(1), 25-38.
By ne, B. M. (2016). S uc u al Equa ion Modeling Wi h AMOS: Basic Concep s, Applica ions,
and P og amming (Thi d ed.): Rou ledge.
Çaka , K. (2020). Tou ophobia: ea o a el esul ing om man-made o na u al disas e s.
Tou ism Re iew.
Can ado , I., & Cas ells, P. (2012). G oup ecommende sys ems: new pe spec i es in he
social web. In Recommende sys ems o he social web (pp. 139-157): Sp inge .
Can ado , I., & Fe nández-Tobías, I. (2014, 2014). On he exploi a ion o use pe sonali y in
ecommende sys ems. Pape p esen ed a he CEUR Wo kshop P oceedings.
Can ado , I., Fe nández-Tobías, I., & Bellogín, A. (2013, 2013). Rela ing pe sonali y ypes wi h
use p e e ences in mul iple en e ainmen domains. Pape p esen ed a he CEUR
wo kshop p oceedings.
Ca nei o, J., Al es, P., Ma ei os, G., & No ais, P. (2019). A Mul i-agen Sys em F amewo k o
Dialogue Games in he G oup Decision-Making Con ex . Pape p esen ed a he Wo ld
Con e ence on In o ma ion Sys ems and Technologies.
Ca nei o, J., And ade, R., Al es, P., Conceição, L., No ais, P., & Ma ei os, G. (2020, 2020). A
consensus-based g oup decision suppo sys em using a mul i-agen Mic oSe ices
app oach. Pape p esen ed a he P oceedings o he 19 h In e na ional Con e ence
on Au onomous Agen s and Mul iAgen Sys ems.
Ca nei o, J., Ma inho, D., Al es, P., Conceição, L., Ma ei os, G., & No ais, P. (2020). A mul iple
c i e ia decision analysis amewo k o dispe sed g oup decision-making con ex s.
Applied Sciences, 10(13), 4614.
Ca nei o, J., Ma inho, D., Ma ei os, G., Jimenez, A., & No ais, P. (2018). Dynamic
a gumen a ion in UbiGDSS. Knowledge and In o ma ion Sys ems, 55(3), 633-669.
Re e ences
284
Ca nei o, J., Ma inho, D., Ma ei os, G., & No ais, P. (2018). A guing wi h Beha io In luence:
A Model o Web-based G oup Decision Suppo Sys ems. In e na ional Jou nal o
In o ma ion Technology & Decision Making, 1-37.
Ca nei o, J., Sa ai a, P., Conceição, L., San os, R., Ma ei os, G., & No ais, P. (2019). P edic ing
sa is ac ion: Pe cei ed decision quali y by decision-make s in web-based g oup
decision suppo sys ems. Neu ocompu ing, 338, 399-417.
Ca nei o, J., Sa ai a, P., Ma inho, D., Ma ei os, G., & No ais, P. (2018). Rep esen ing
decision-make s using s yles o beha io : An app oach designed o g oup decision
suppo sys ems. Cogni i e Sys ems Resea ch, 47, 109-132.
Ca alho, L. d. F., Pianowski, G., & Gonçal es, A. P. (2020). Pe sonali y di e ences and COVID-
19: a e ex o e sion and conscien iousness pe sonali y ai s associa ed wi h
engagemen wi h con ainmen measu es? T ends in psychia y and psycho he apy, 42,
179-184.
Cas ells, P., Hu ley, N., & Va gas, S. (2021). No el y and di e si y in ecommende sys ems. In
Recommende sys ems handbook (pp. 603-646): Sp inge .
Cas ells, P., & Jannach, D. (2024). Recommende Sys ems. In In o ma ion Re ie al: Ad anced
Topics and Techniques (pp. 49-109).
Cas o, J., Quesada, F. J., Paloma es, I., & Ma inez, L. (2015). A Consensus‐D i en G oup
Recommende Sys em. In e na ional Jou nal o In elligen Sys ems, 30(8), 887-906.
Cecca oni, L., Codina, V., Palau, M., & Pous, M. (2009). PaTac: U ban, ubiqui ous, pe sonalized
se ices o ci izens and ou is s. Pape p esen ed a he 2009 Thi d In e na ional
Con e ence on Digi al Socie y.
Celli, F., B uni, E., & Lep i, B. (2014). Au oma ic pe sonali y and in e ac ion s yle ecogni ion
om acebook p o ile pic u es. Pape p esen ed a he P oceedings o he 22nd ACM
in e na ional con e ence on Mul imedia.
Chand amouli, B., Le andoski, J. J., Eldawy, A., & Mokbel, M. F. (2011). S eam ec: a eal- ime
ecommende sys em. Pape p esen ed a he P oceedings o he 2011 ACM SIGMOD
In e na ional Con e ence on Managemen o da a.
Chen, L., Wu, W., & He, L. (2016). Pe sonali y and ecommenda ion di e si y. In Emo ions and
Pe sonali y in Pe sonalized Se ices (pp. 201-225): Sp inge .
Chen, Y., & Pu, P. (2012). CoFeel: Using emo ions o social in e ac ion in g oup ecommende
sys ems. Pape p esen ed a he Fi s In e na ional Wo kshop on Recommenda ion
Technologies o Li es yle Change (LIFESTYLE 2012).
Cheng, C., & Chau, C.-l. (2022). Gami ica ion-based in e en ion o enhancing eam
e ec i eness and coping lexibili y: Randomized con olled ial. F on ie s in
psychia y, 13, 941252.
Chi a anjan, G., Blom, J., & Ga ica-Pe ez, D. (2013). Mining la ge-scale sma phone da a o
pe sonali y s udies. Pe sonal and Ubiqui ous Compu ing, 17(3), 433-450.
Cla k, L. A., & Wa son, D. (2016). Cons uc ing alidi y: Basic issues in objec i e scale
de elopmen .
Cohen, E. (1972). Towa d a sociology o in e na ional ou ism. Social esea ch, 164-182.
Cohen, E. (1979). Re hinking he sociology o ou ism. Annals o ou ism esea ch, 6(1), 18-
35.
Cohen, J. (2013). S a is ical powe analysis o he beha io al sciences: Academic p ess.
Collins-K eine , N., & Klio , N. (2000). Pilg image ou ism in he Holy Land: The beha iou al
cha ac e is ics o Ch is ian pilg ims. GeoJou nal, 50(1), 55-67.
Coope , S., Kha ib, F., T euille, A., Ba be o, J., Lee, J., Beenen, M., . . . Popo ić, Z. (2010).
P edic ing p o ein s uc u es wi h a mul iplaye online game. Na u e, 466(7307),
756-760.
Re e ences
285
Cos a J , P. T., McC ae, R. R., & Kay, G. G. (1995). Pe sons, places, and pe sonali y: Ca ee
assessmen using he Re ised NEO Pe sonali y In en o y. Jou nal o Ca ee
Assessmen , 3(2), 123-139.
Cos a, P. T., & MacC ae, R. R. (1992). Re ised NEO pe sonali y in en o y (NEO PI-R) and NEO
i e- ac o in en o y (NEO-FFI): P o essional manual: Psychological Assessmen
Resou ces, Inco po a ed.
Cos a, P. T., & McC ae, R. R. (1988). Pe sonali y in adul hood: a six-yea longi udinal s udy o
sel - epo s and spouse a ings on he NEO Pe sonali y In en o y. Jou nal o
pe sonali y and social psychology, 54(5), 853.
Co ing on, P., Adams, J., & Sa gin, E. (2016). Deep neu al ne wo ks o you ube
ecommenda ions. Pape p esen ed a he P oceedings o he 10 h ACM con e ence on
ecommende sys ems.
Cowley, B., & Cha les, D. (2016). Beha le s: a me hod o p ac ical playe modelling using
psychology-based playe ai s and domain speci ic ea u es. Use Modeling and Use -
Adap ed In e ac ion, 26, 257-306.
C omp on, J. L. (1979). Mo i a ions o pleasu e aca ion. Annals o ou ism esea ch, 6(4),
408-424.
C o s, J. C. (1990). Towa ds an unde s anding o he no el y seeking d i e in pleasu e
a ele s. Visions in Leisu e and Business, 9(3), 9.
C owne, D. P., & Ma lowe, D. (1960). Ma lowe-C owne social desi abili y scale. Jou nal o
Consul ing Psychology.
da C uz, J. A., Du ao, F. A., & Rosse i, R. J. (2024). E alua ing Di e si ica ion in G oup
Recommenda ion o Poin s o In e es . Pape p esen ed a he 20 h In e na ional
Con e ence on Web In o ma ion Sys ems and Technologies (WEBIST 2024), Po o,
Po ugal.
Dann, G. M. (1977). Anomie, ego-enhancemen and ou ism. Annals o ou ism esea ch, 4(4),
184-194.
Da a, S., Chowda y, C. R., & Kuma , C. (2020). A su ey on g oup ecommende sys ems.
Jou nal o in elligen in o ma ion sys ems, 54(2), 271-295.
de CA Zieseme , A., Mülle , L., & Sil ei a, M. S. (2014). Jus a e i ! Gami ica ion as pa o
ecommenda ion. Pape p esen ed a he In e na ional Con e ence on Human-
Compu e In e ac ion.
De Campos, L. M., Fe nández-Luna, J. M., & Hue e, J. F. (2024). An explainable con en -based
app oach o ecommende sys ems: a case s udy in jou nal ecommenda ion o
pape submission. Use Modeling and Use -Adap ed In e ac ion, 1-35.
de Ca alho, A. P., B az, C. S., dos San os, S. M., Fe ei a, R. A., & P a es, R. O. (2024). Se ious
games o child en wi h au ism spec um diso de : A sys ema ic li e a u e e iew.
In e na ional Jou nal o Human–Compu e In e ac ion, 40(14), 3655-3682.
del Ca men Rod íguez-He nández, M., Ila i, S., He moso, R., & T illo-Lado, R. (2017).
Towa ds ajec o y-based ecommenda ions in museums: e alua ion o s a egies
using mixed syn he ic and eal da a. P ocedia Compu e Science, 113, 234-239.
Deldjoo, Y., He, Z., McAuley, J., Ko iko , A., Sanne , S., Ramisa, A., . . . Milano, S. (2024). A
e iew o mode n ecommende sys ems using gene a i e models (gen- ecsys). Pape
p esen ed a he P oceedings o he 30 h ACM SIGKDD Con e ence on Knowledge
Disco e y and Da a Mining.
Delic, A., Emamgholizadeh, H., Nguyen, T. N., & Ricci, F. (2024). CHARM: a G oup Decision
Making Suppo Cha bo . Pape p esen ed a he Companion P oceedings o he 29 h
In e na ional Con e ence on In elligen Use In e aces.
Re e ences
286
Delic, A., & Mas ho , J. (2018). G oup Recommende Sys ems. Pape p esen ed a he
P oceedings o he 26 h Con e ence on Use Modeling, Adap a ion and
Pe sonaliza ion.
Delic, A., Neidha d , J., Nguyen, N., & Ricci, F. (2016). Resea ch Me hods o G oup
Recommende Sys em. In: CEUR-WS.
Delic, A., Neidha d , J., Nguyen, T. N., & Ricci, F. (2018). An obse a ional use s udy o g oup
ecommende sys ems in he ou ism domain. In o ma ion Technology & Tou ism, 19,
87-116.
Delic, A., Neidha d , J., Nguyen, T. N., Ricci, F., Rook, L., We hne , H., & Zanke , M. (2016).
Obse ing g oup decision making p ocesses. Pape p esen ed a he P oceedings o he
10 h ACM con e ence on ecommende sys ems.
Delic, A., Neidha d , J., Rook, L., We hne , H., & Zanke , M. (2017). Resea ching indi idual
sa is ac ion wi h g oup decisions in ou ism: expe imen al e idence. Pape p esen ed a
he In o ma ion and Communica ion Technologies in Tou ism 2017: P oceedings o
he In e na ional Con e ence in Rome, I aly, Janua y 24-26, 2017.
Delic, A., Neidha d , J., & We hne , H. (2016). A e sun lo e s ne ous. Pape p esen ed a he
Resea ch no e a en e 2016 e ou ism con e ence. Bilbao, Spain.
Delić, A., Nguyen, T. N., & Tkalčič, M. (2020). G oup decision-making and designing g oup
ecommende sys ems. Handbook o e-Tou ism, 1-23.
Dhelim, S., Aung, N., Bou as, M. A., Ning, H., & Camb ia, E. (2021). A su ey on pe sonali y-
awa e ecommenda ion sys ems. A i icial In elligence Re iew, 1-46.
Dhelim, S., Chen, L., Aung, N., Zhang, W., & Ning, H. (2023). A hyb id pe sonali y-awa e
ecommenda ion sys em based on pe sonali y ai s and ypes models. Jou nal o
Ambien In elligence and Humanized Compu ing, 14(9), 12775-12788.
Dhelim, S., Ning, H., Aung, N., Huang, R., & Ma, J. (2020). Pe sonali y-awa e p oduc
ecommenda ion sys em based on use in e es s mining and me apa h disco e y.
IEEE T ansac ions on Compu a ional Social Sys ems, 8(1), 86-98.
Diamond, L., Tondello, G. F., Ma czewski, A., Nacke, L. E., & Tscheligi, M. (2015). The HEXAD
gami ica ion use ypes ques ionnai e: Backg ound and de elopmen p ocess. Pape
p esen ed a he CHI PLAY 2015-The ACM SIGCHI Annual Symposium on Compu e -
Human In e ac ion in Play.
Dide o , D., & d'Alembe , J. L. R. (1776). Encyclopédie, ou, Dic ionnai e aisonné des sciences,
des a s e des mé ie s (Vol. 5): Pe gamon P ess.
Digman, J. M. (1990). Pe sonali y s uc u e: Eme gence o he i e- ac o model. Annual
e iew o psychology, 41(1), 417-440.
DiS e ano, C., Zhu, M., & Mind ila, D. (2009). Unde s anding and using ac o sco es:
Conside a ions o he applied esea che . P ac ical Assessmen , Resea ch, and
E alua ion, 14(1), 20.
Djaou i, D., Al a ez, J., & Jessel, J.-P. (2011). Classi ying se ious games: he G/P/S model. In
Handbook o esea ch on imp o ing lea ning and mo i a ion h ough educa ional
games: Mul idisciplina y app oaches (pp. 118-136): IGI global.
Djebali, S., Gabo , Q., & Gue a d, G. (2023). Hie a chical Clus e ing and Measu e o Tou ism
P o iling. Pape p esen ed a he Web and Big Da a: 6 h In e na ional Join
Con e ence, APWeb-WAIM 2022, Nanjing, China, No embe 25–27, 2022,
P oceedings, Pa II.
Dolnica , S. (2005). Unde s anding ba ie s o leisu e a el: Tou is ea s as a ma ke ing
basis. Jou nal o Vaca ion Ma ke ing, 11(3), 197-208.
Do i, A., Kanhe e, S. S., & Ju dak, R. (2018). Mul i-agen sys ems: A su ey. IEEE Access, 6,
28573-28593.

Re e ences
287
Dueñas-Le ín, J., La a-Cab e a, R., O ega, F., & Bobadilla, J. (2023). Deep Neu al Agg ega ion
o Recommending I ems o G oup o Use s. a Xi p ep in a Xi :2307.09447.
Dumon , F. (2010). A his o y o pe sonali y psychology: Theo y, science, and esea ch om
Hellenism o he wen y- i s cen u y: Camb idge uni e si y p ess.
Dunn, G., Wie sema, J., Ham, J., & A oyo, L. (2009, 2009). E alua ing in e ace a ian s on
pe sonali y acquisi ion o ecommende sys ems. Pape p esen ed a he In e na ional
Con e ence on Use Modeling, Adap a ion, and Pe sonaliza ion.
Eachus, P. (2004). Using he B ie Sensa ion Seeking Scale (BSSS) o p edic holiday
p e e ences. Pe sonali y and indi idual di e ences, 36(1), 141-153.
El Majjodi, A., Elahi, M., El Ioini, N., & T a ne , C. (2020). Towa ds Gene a ing Pe sonalized
Coun y Recommenda ion. Pape p esen ed a he Adjunc Publica ion o he 28 h
ACM Con e ence on Use Modeling, Adap a ion and Pe sonaliza ion.
Eldeswky, K., Elazab, F., Bolock, A. E., & Abdennadhe , S. (2023). Cha ac e -Based Habi
Recommende Sys em. Pape p esen ed a he In e na ional Con e ence on P ac ical
Applica ions o Agen s and Mul i-Agen Sys ems.
Ellio , A. J., & McG ego , H. A. (2001). A 2× 2 achie emen goal amewo k. Jou nal o
pe sonali y and social psychology, 80(3), 501.
Emamgholizadeh, H., Delić, A., & Ricci, F. (2024). P e e ence Managemen in a Single Use
G oup Recommende Sys em. Pape p esen ed a he ENTER e-Tou ism Con e ence.
Eysenck, H., & Rein, M. (1998). Dimensions o pe sonali y: Rou ledge.
Eysenck, H. J. (1958). The scien i ic s udy o pe sonali y: Rou ledge & Paul.
Eysenck, H. J. (1966). Pe sonali y and expe imen al psychology. Bulle in o he B i ish
Psychological Socie y.
Eysenck, H. J. (1982). Pe sonali y, gene ics, and beha io : Selec ed pape s.
Ezugwu, A. E., Iko un, A. M., Oyelade, O. O., Abualigah, L., Agushaka, J. O., Eke, C. I., & Akinyelu,
A. A. (2022). A comp ehensi e su ey o clus e ing algo i hms: S a e-o - he-a
machine lea ning applica ions, axonomy, challenges, and u u e esea ch p ospec s.
Enginee ing applica ions o a i icial in elligence, 110, 104743.
Fab iga , L. R., Wegene , D. T., MacCallum, R. C., & S ahan, E. J. (1999). E alua ing he use o
explo a o y ac o analysis in psychological esea ch. Psychological me hods, 4(3),
272.
Faullan , R., Ma zle , K., & Moo adian, T. A. (2011). Pe sonali y, basic emo ions, and
sa is ac ion: P ima y emo ions in he moun ainee ing expe ience. Tou ism
Managemen , 32(6), 1423-1430.
Feil, S., K e ze , M., We de , K., & Maedche, A. (2016). Using gami ica ion o ackle he cold-
s a p oblem in ecommende sys ems. Pape p esen ed a he P oceedings o he
19 h ACM Con e ence on Compu e Suppo ed Coope a i e Wo k and Social
Compu ing Companion.
Feldman, R., & Sange , J. (2007). The ex mining handbook: ad anced app oaches in analyzing
uns uc u ed da a: Camb idge uni e si y p ess.
Fel e nig, A., A as, M., Helic, D., T an, T. N. T., S e inge , M., & Same , R. (2023). Algo i hms
o g oup ecommenda ion. In G oup ecommende sys ems: An in oduc ion (pp. 29-
61): Sp inge .
Fe we da, B., Schedl, M., & Tkalcic, M. (2015). P edic ing pe sonali y ai s wi h ins ag am
pic u es. Pape p esen ed a he P oceedings o he 3 d Wo kshop on Emo ions and
Pe sonali y in Pe sonalized Sys ems 2015.
Field, A. (2013). Disco e ing s a is ics using IBM SPSS s a is ics: sage.
Finne y, A. N., Lep i, B., & Pianesi, F. (2016). Acquisi ion o pe sonali y. In Emo ions and
Pe sonali y in Pe sonalized Se ices (pp. 81-99): Sp inge .
Re e ences
288
Fleische , A., & Pizam, A. (2002). Tou ism cons ain s among Is aeli senio s. Annals o
ou ism esea ch, 29(1), 106-123.
Floyd, M. F., Gibson, H., Penning on-G ay, L., & Thapa, B. (2004). The e ec o isk pe cep ions
on in en ions o a el in he a e ma h o Sep embe 11, 2001. Jou nal o T a el &
Tou ism Ma ke ing, 15(2-3), 19-38.
Foead, D., Ghi a i, A., Kusuma, M. B., Hana iah, N., & Gunawan, E. (2021). A sys ema ic
li e a u e e iew o A* pa h inding. P ocedia Compu e Science, 179, 507-514.
Fo sy h, D. R. (2018). G oup dynamics: Cengage Lea ning.
F iedman, H. S., & Schus ack, M. W. (1999). Pe sonali y: Classic heo ies and mode n esea ch:
Allyn and Bacon Bos on, MA.
Gao, J., & Ke s e e , D. L. (2016). Using an in e sec ionali y pe spec i e o unco e olde
Chinese emale's pe cei ed a el cons ain s and nego ia ion s a egies. Tou ism
Managemen , 57, 128-138.
Ga cia, I., Sebas ia, L., & Onaindia, E. (2011). On he design o indi idual and g oup
ecommende sys ems o ou ism. Expe sys ems wi h applica ions, 38(6), 7683-
7692.
Ga cia, I., Sebas ia, L., Onaindia, E., & Guzman, C. (2009). A g oup ecommende sys em o
ou is ac i i ies. Pape p esen ed a he In e na ional con e ence on elec onic
comme ce and web echnologies.
Ga alas, D., & Ken e is, M. (2011). A web-based pe asi e ecommenda ion sys em o
mobile ou is guides. Pe sonal and Ubiqui ous Compu ing, 15(7), 759-770.
Ga alas, D., Kons an opoulos, C., Mas akas, K., & Pan ziou, G. (2014). Mobile ecommende
sys ems in ou ism. Jou nal o ne wo k and compu e applica ions, 39, 319-333.
Gee, C. Y., Choy, D. J., & Makens, J. C. (1984). The a el indus y: AVI Publishing Company, Inc.
Geo ge, D., & Malle y, P. (2019). IBM SPSS s a is ics 26 s ep by s ep: A simple guide and
e e ence: Rou ledge.
Gibson, H., & Yiannakis, A. (2002). Tou is oles: Needs and he li ecou se. Annals o ou ism
esea ch, 29(2), 358-383.
Gimmon, E., Rob, O., Fa ja, Y., & Golan, M. (2023). The challenge o physical puzzle games:
Exci emen o achie emen ? In Ad ances and Lessons in Spo s (pp. 321-332):
Newcas le upon Tyne, UK: Camb idge Schola s Publishing.
Gladwell, N. J., & Bedini, L. A. (2004). In sea ch o los leisu e: The impac o ca egi ing on
leisu e a el. Tou ism Managemen , 25(6), 685-693.
Gno h, J. (1997). Tou ism mo i a ion and expec a ion o ma ion. Annals o ou ism esea ch,
24(2), 283-304.
Godo ykh, M., Pizam, A., & Bahja, F. (2021). An eceden s and ou comes o heal h isk
pe cep ions in ou ism, ollowing he COVID-19 pandemic. Tou ism Re iew.
Golbeck, J., Robles, C., Edmondson, M., & Tu ne , K. (2011). P edic ing pe sonali y om
wi e . Pape p esen ed a he 2011 IEEE hi d in e na ional con e ence on p i acy,
secu i y, isk and us and 2011 IEEE hi d in e na ional con e ence on social
compu ing.
Goldbe g, D., Nichols, D., Oki, B. M., & Te y, D. (1992). Using collabo a i e il e ing o wea e
an in o ma ion apes y. Communica ions o he ACM, 35(12), 61-70.
Goldbe g, L. R. (1971). A his o ical su ey o pe sonali y scales and in en o ies: O egon
Resea ch Ins i u e.
Goldbe g, L. R. (1990). An al e na i e" desc ip ion o pe sonali y": he big- i e ac o
s uc u e. Jou nal o pe sonali y and social psychology, 59(6), 1216.
Golds ein, G., & He sen, M. (2000). Handbook o psychological assessmen : Else ie .
Gosling, S. D., Ren ow, P. J., & Swann J , W. B. (2003). A e y b ie measu e o he Big-Fi e
pe sonali y domains. Jou nal o esea ch in pe sonali y, 37(6), 504-528.
Re e ences
289
Gosling, S. D., Vazi e, S., S i as a a, S., & John, O. P. (2004). Should we us web-based
s udies? A compa a i e analysis o six p econcep ions abou in e ne ques ionnai es.
Ame ican psychologis , 59(2), 93.
G a e e , F. J., Wallnau, L. B., Fo zano, L.-A. B., & Wi naue , J. E. (2020). Essen ials o s a is ics
o he beha io al sciences: Cengage Lea ning.
G ay, H. P. (1970). In e na ional a el--in e na ional ade: Hea h Lexing on Books.
G e zel, U., Mi sche, N., Hwang, Y.-H., & Fesenmaie , D. R. (2004). Tell me who you a e and I
will ell you whe e o go: Use o a el pe sonali ies in des ina ion ecommenda ion
sys ems. In o ma ion Technology & Tou ism, 7(1), 3-12.
G e zel, U., Mi sche, N., Hwang, Y.-H., & Fesenmaie , D. R. (2006). T a el pe sonali y es ing
o des ina ion ecommenda ion sys ems. Des in Recomm Sys Beha Found Appl.
CABI, Ox o dshi e, 121-136.
G i i h, D. A., & Albanese, P. J. (1996). An examina ion o Plog's psychog aphic a el model
wi hin a s uden popula ion. Jou nal o T a el Resea ch, 34(4), 47-51.
Gus a ii, B. (2012). How o p epa e a scien i ic doc o al disse a ion based on esea ch
a icles: Camb idge Uni e si y P ess.
Guzzi, F., Ricci, F., & Bu ke, R. (2011). In e ac i e mul i-pa y c i iquing o g oup
ecommenda ion. Pape p esen ed a he P oceedings o he i h ACM con e ence on
Recommende sys ems.
Hai , J. F., Ande son, R., Ta ham, R., & Black, W. (2009). Mul i a ia e da a analysis: A Global
Pe spec i e (7 h edi ion ed.): Uppe Saddle Ri e : P en ice Hall.
Haizel, P., Ve nanda, G., Wawolangi, K. A., & Hana iah, N. (2021). Pe sonali y assessmen
ideo game based on he i e- ac o model. P ocedia Compu e Science, 179, 566-573.
Hakulinen, L., Au inen, T., & Ko honen, A. (2015). The e ec o achie emen badges on
s uden s’ beha io : An empi ical s udy in a uni e si y-le el compu e science cou se.
In e na ional Jou nal o Eme ging Technologies in Lea ning (iJET), 10(1), 18-29.
Halim, Z., A i , M., Rashid, A., & Edwin, C. A. (2017). P o iling playe s using eal-wo ld
da ase s: clus e ing he da a and co ela ing he esul s wi h he big- i e pe sonali y
ai s. IEEE T ansac ions on A ec i e Compu ing, 10(4), 568-584.
Hama i, J. (2013). T ans o ming homo economicus in o homo ludens: A ield expe imen on
gami ica ion in a u ili a ian pee - o-pee ading se ice. Elec onic comme ce
esea ch and applica ions, 12(4), 236-245.
Hama i, J., Koi is o, J., & Sa sa, H. (2014). Does gami ica ion wo k?--a li e a u e e iew o
empi ical s udies on gami ica ion. Pape p esen ed a he 2014 47 h Hawaii
in e na ional con e ence on sys em sciences (HICSS).
Hama i, J., She no , D. J., Rowe, E., Colle , B., Asbell-Cla ke, J., & Edwa ds, T. (2016).
Challenging games help s uden s lea n: An empi ical s udy on engagemen , low and
imme sion in game-based lea ning. Compu e s in Human Beha io , 54, 170-179.
Hassani, A., & Mogha emi, S. (2019). Muslims’ a el mo i a ions and a el p e e ences:
The impac o mo i a ional ac o s on Islamic se ice, hedonic and p oduc
p e e ences. Jou nal o Islamic Ma ke ing.
Ha haway, S. R., & McKinley, J. C. (1951). Minneso a mul iphasic pe sonali y in en o y;
manual, e ised.
He, B., He, X., Zhang, R., Zhang, Y., Tang, R., & Ma, C. (2023). Dynamic Embedding Size Sea ch
wi h Minimum Reg e o S eaming Recommende Sys em. Pape p esen ed a he
P oceedings o he 32nd ACM In e na ional Con e ence on In o ma ion and
Knowledge Managemen .
Hee e , C., Lee, Y.-H., Medle , B., & Mage ko, B. (2011). Beyond playe ypes: gaming
achie emen goal. Pape p esen ed a he P oceedings o he 2011 ACM SIGGRAPH
symposium on ideo games.
Re e ences
290
Hei mann, S. (2011). Tou is beha iou and ou ism mo i a ion. Resea ch hemes o ou ism,
31-44.
He locke , J. L., Kons an, J. A., & Riedl, J. (2000). Explaining collabo a i e il e ing
ecommenda ions. Pape p esen ed a he P oceedings o he 2000 ACM con e ence
on Compu e suppo ed coope a i e wo k.
He e a-Viedma, E., Chiclana, F., He e a, F., & Alonso, S. (2007). G oup decision-making
model wi h incomple e uzzy p e e ence ela ions based on addi i e consis ency.
IEEE T ansac ions on Sys ems, Man, and Cybe ne ics, Pa B (Cybe ne ics), 37(1), 176-
189.
He z, M. R. (1936). The me hod o adminis a ion o he Ro schach ink-blo es . Child
De elopmen , 237-254.
Heung, V. C., & Leong, J. S. (2006). T a el demand and beha io o uni e si y s uden s in
Hong Kong. Asia Paci ic Jou nal o Tou ism Resea ch, 11(1), 81-96.
He ne , A., Ma ch, S. T., Pa k, J., & Ram, S. (2004). Design science esea ch in in o ma ion
sys ems. MIS qua e ly, 28(1), 75-105.
Hi sh, J. B. (2010). Pe sonali y and en i onmen al conce n. Jou nal o En i onmen al
Psychology, 30(2), 245-248.
Hol g a es, T. (2011). Tex messaging, pe sonali y, and he social con ex . Jou nal o esea ch
in pe sonali y, 45(1), 92-99.
Hox e , A. L., & Les e , D. (1988). Tou is beha io and pe sonali y. Pe sonali y and indi idual
di e ences, 9(1), 177-178.
Hoyle, R. H., S ephenson, M. T., Palmg een, P., Lo ch, E. P., & Donohew, R. L. (2002). Reliabili y
and alidi y o a b ie measu e o sensa ion seeking. Pe sonali y and indi idual
di e ences, 32(3), 401-414.
Hu, R., & Pu, P. (2009). A compa a i e use s udy on a ing s. pe sonali y quiz based
p e e ence elici a ion me hods. Pape p esen ed a he P oceedings o he 14 h
in e na ional con e ence on In elligen use in e aces.
Huang, L., & Tsai, H.-T. (2003). The s udy o senio a ele beha io in Taiwan. Tou ism
Managemen , 24(5), 561-574.
Hung, K., Bai, X., & Lu, J. (2016). Unde s anding a el cons ain s among he elde ly in Hong
Kong: A compa a i e s udy o he elde ly li ing in p i a e and in public housing.
Jou nal o T a el & Tou ism Ma ke ing, 33(7), 1051-1070.
Hun e ‐Jones, P., & Blackbu n, A. (2007). Unde s anding he ela ionship be ween holiday
aking and sel ‐assessed heal h: an explo a o y s udy o senio ou ism. In e na ional
Jou nal o Consume S udies, 31(5), 509-516.
Ismailoglu, F. (2022). Agg ega ing use p e e ences in g oup ecommende sys ems: A
c owdsou cing app oach. Decision Suppo Sys ems, 152, 113663.
Ismailo , A. (2017). The ela ion be ween pe sonali y ai s and psychog aphic posi ions o
a el des ina ions. Uni e si y o S a ange , No way,
Iso-Ahola, S. E. (1982). Towa d a social psychological heo y o ou ism mo i a ion: A
ejoinde . Annals o ou ism esea ch, 9(2), 256-262.
Jackson, M., & Inbaka an, R. (2006). De elopmen o pe sonali y ypes o p edic ou is
beha iou owa ds ou is p oduc consump ion. Tou is s and ou ism, 1-1.
Jackson, M., Whi e, G., & Whi e, M. G. (2001). De eloping a ou is pe sonali y ypology. Pape
p esen ed a he CAUTHE 2001: Capi alising on Resea ch; P oceedings o he 11 h
Aus alian Tou ism and Hospi ali y Resea ch Con e ence.
Jackson, M. S., & Inbaka an, R. (2006). De elopmen o a ou is pe sonali y in en o y.
CAUTHE 2006: To he ci y and beyond, 932.
Jackson, M. S., Schmie e , C. L., & Whi e, G. N. (1999). Is he e a unique ou is pe sonali y
which is p edic i e o ou is beha iou ? Pape p esen ed a he CAUTHE 1999:
Re e ences
297
Expe imen al S udy. Pape p esen ed a he 2024 IEEE In e na ional Con e ence on
Ad anced Lea ning Technologies (ICALT).
O oo, F. E., & Kim, S. (2020). Analysis o s udies on he a el mo i a ions o senio ou is s
om 1980 o 2017: p og ess and u u e di ec ions. Cu en issues in ou ism, 23(4),
393-417.
O oo, F. E., Kim, S., Ag usa, J., & Lema, J. (2021). Classi ica ion o senio ou is s acco ding o
pe sonali y ai s. Asia Paci ic Jou nal o Tou ism Resea ch, 26(5), 539-556.
O oo, F. E., Kim, S., & Pa k, J. (2020). Mo i a ion-based segmen a ion o Chinese senio
a ele s: The ole o p e e ences, sociodemog aphic, and a el- ela ed ea u es.
Jou nal o Vaca ion Ma ke ing, 26(4), 457-472.
Pal, C.-V., Leon, F., Pap zycki, M., & Ganzha, M. (2020). A e iew o pla o ms o he
de elopmen o agen sys ems. a Xi p ep in a Xi :2007.08961.
Palomäki, J., Laakasuo, M., Cas én, S., Saas amoinen, J., Kainulainen, T., & Suhonen, N. (2021).
Online be ing in ensi y is linked wi h Ex a e sion and Conscien iousness. Jou nal o
pe sonali y, 89(5), 1081-1094.
Pa k, D.-B., & Yoon, Y.-S. (2009). Segmen a ion by mo i a ion in u al ou ism: A Ko ean case
s udy. Tou ism Managemen , 30(1), 99-108.
Pa k, M.-H., Pa k, H.-S., & Cho, S.-B. (2008). Res au an ecommenda ion o g oup o people in
mobile en i onmen s using p obabilis ic mul i-c i e ia decision making. Pape
p esen ed a he Compu e -Human In e ac ion: 8 h Asia-Paci ic Con e ence, APCHI
2008 Seoul, Ko ea, July 6-9, 2008 P oceedings 8.
Passa a o, P., Cini, F., Boi, L., D'Angelo, M., Hee ing, M. S., Luche i, L., . . . Pa izi, F. (2015). The
“sus ainable ou is ”: Values, a i udes, and pe sonali y ai s. Tou ism and Hospi ali y
Resea ch, 15(4), 225-239.
Pa uelli, R., & Nijkamp, P. (2016). T a el mo i a ions o senio s: a e iew and a me a-
analy ical assessmen . Tou ism economics, 22(4), 847-862.
Pazzani, M. J. (1999). A amewo k o collabo a i e, con en -based and demog aphic
il e ing. A i icial In elligence Re iew, 13, 393-408.
Pea ce, P. L. (1993). Fundamen als o ou is mo i a ion. Tou ism esea ch: C i iques and
challenges, 85-105.
Pea ce, P. L., & Cal abiano, M. L. (1983). In e ing a el mo i a ion om a ele s'
expe iences. Jou nal o T a el Resea ch, 22(2), 16-20.
Pea ce, P. L., & Lee, U.-I. (2005). De eloping he a el ca ee app oach o ou is mo i a ion.
Jou nal o T a el Resea ch, 43(3), 226-237.
Ped egon, C. A., Fa ley, R. L., Da is, A., Wood, J. M., & Cla k, R. D. (2012). Social desi abili y,
pe sonali y ques ionnai es, and he “be e han a e age” e ec . Pe sonali y and
indi idual di e ences, 52(2), 213-217.
Pe e s, K., Tuunanen, T., Ro henbe ge , M. A., & Cha e jee, S. (2007). A design science
esea ch me hodology o in o ma ion sys ems esea ch. Jou nal o managemen
in o ma ion sys ems, 24(3), 45-77.
Pei, F., Gao, Y., Yan, A., Zhou, M., & Wu, J. (2024). Con lic elimina ion based on opinion
dynamics in uzzy g oup decision-making. Expe sys ems wi h applica ions, 124308.
Pe ik, E., De Ruy e , B., Ma kopoulos, P., & Eggen, B. (2004). The sensi i i ies o use p o ile
in o ma ion in music ecommende sys ems. P oceedings o P i a e, Secu i y, T us ,
137-141.
Pes ana, M. H., & Gagei o, J. N. (2008). Análise de dados pa a ciências sociais: a
complemen a idade do SPSS.
Pham, D. T., Dimo , S. S., & Nguyen, C. (2004). An inc emen al K-means algo i hm.
P oceedings o he Ins i u ion o Mechanical Enginee s, Pa C: Jou nal o Mechanical
Enginee ing Science, 218(7), 783-795.

Re e ences
298
Pine, B. J., Pine, J., & Gilmo e, J. H. (1999). The expe ience economy: wo k is hea e & e e y
business a s age: Ha a d Business P ess.
Ping, Y., Li, Y., & Zhu, J. (2024). Beyond accu acy measu es: he e ec o di e si y, no el y and
se endipi y in ecommende sys ems on use engagemen . Elec onic Comme ce
Resea ch, 1-28.
Plog, S. (2001). Why des ina ion a eas ise and all in popula i y: An upda e o a Co nell
Qua e ly classic. Co nell ho el and es au an adminis a ion qua e ly, 42(3), 13-24.
Plog, S. C. (1974). Why des ina ion a eas ise and all in popula i y. Co nell ho el and
es au an adminis a ion qua e ly, 14(4), 55-58.
Plog, S. C. (1991). Leisu e a el: making i a g ow h ma ke .... again! : John Wiley and Sons,
Inc.
Plog, S. C. (1994). De eloping and using psychog aphics in ou ism esea ch. T a el, ou ism
and hospi ali y esea ch, 209-231.
Plog, S. C. (2002). The powe o psychog aphics and he concep o en u esomeness. Jou nal
o T a el Resea ch, 40(3), 244-251.
Poon, A., & Adams, E. (2000). How he B i ish will a el 2005. Biele eld: Tou ism In elligence
In e na ional.
Poon, K. Y., & Huang, W.-J. (2017). Pas expe ience, a ele pe sonali y and ipog aphics on
in en ion o use Ai bnb. In e na ional Jou nal o Con empo a y Hospi ali y
Managemen , 29(9), 2425-2443.
Pyle, D. (1999). Da a p epa a ion o da a mining: mo gan kau mann.
Quwaide , M., Alabed, A., & Duwai i, R. (2023). Shoo e ideo games o pe sonali y
p edic ion using i e ac o model ai s and machine lea ning. Simula ion Modelling
P ac ice and Theo y, 122, 102665.
Rami es, A., B andao, F., & Sousa, A. C. (2018). Mo i a ion-based clus e analysis o
in e na ional ou is s isi ing a Wo ld He i age Ci y: The case o Po o, Po ugal.
Jou nal o Des ina ion Ma ke ing & Managemen , 8, 49-60.
Rashid, A. M., Albe , I., Cosley, D., Lam, S. K., McNee, S. M., Kons an, J. A., & Riedl, J. (2002).
Ge ing o know you: lea ning new use p e e ences in ecommende sys ems. Pape
p esen ed a he P oceedings o he 7 h in e na ional con e ence on In elligen use
in e aces.
Ra i, L., De a ajan, M., Sangaiah, A. K., Wang, L., & Sub amaniyaswamy, V. (2021). An
in elligen loca ion ecommende sys em u ilising mul i-agen induced cogni i e
beha iou al model. En e p ise In o ma ion Sys ems, 15(10), 1376-1394.
Rawlings, D., & Cianca elli, V. (1997). Music p e e ence and he i e- ac o model o he NEO
Pe sonali y In en o y. Psychology o Music, 25(2), 120-132.
RecSys. (2023). 17 h ACM Con e ence on Recommende Sys ems. Re ie ed om
h ps:// ecsys.acm.o g/ ecsys23/
Resnick, P., Iaco ou, N., Suchak, M., Be gs om, P., & Riedl, J. (1994). G ouplens: An open
a chi ec u e o collabo a i e il e ing o ne news. Pape p esen ed a he P oceedings
o he 1994 ACM con e ence on Compu e suppo ed coope a i e wo k.
Resnick, P., & Va ian, H. R. (1997). Recommende sys ems. Communica ions o he ACM,
40(3), 56-58.
Ricci, F. (2002). T a el ecommende sys ems. IEEE In elligen Sys ems, 17(6), 55-57.
Ricci, F. (2010). Mobile ecommende sys ems. In o ma ion Technology & Tou ism, 12(3),
205-231.
Ricci, F. (2022). Recommende sys ems in ou ism. In Handbook o e-Tou ism (pp. 457-474):
Sp inge .
Ricci, F., Rokach, L., & Shapi a, B. (2015). Recommende Sys ems Handbook (2nd ed.):
Sp inge .
Re e ences
299
Rich, E. (1979). Use modeling ia s e eo ypes. Cogni i e science, 3(4), 329-354.
Roshchina, A. (2012). TWIN: Pe sonali y-based Recommende Sys em. Ins i u e o Technology
Tallagh , Dublin,
Rousseeuw, P. J. (1987). Silhoue es: a g aphical aid o he in e p e a ion and alida ion o
clus e analysis. Jou nal o compu a ional and applied ma hema ics, 20, 53-65.
Rowe, A. (2015). Imme sion in mixed eali y spaces. No way: The Oslo School o A chi ec u e
and Design.
Sabi , M. (2011). Wea he and a el beha iou . V ije Uni e si ei Ams e dam,
San os, C. P., Hu chinson, K., Khan, V.-J., & Ma kopoulos, P. (2019). P o iling pe sonali y ai s
wi h games. ACM T ansac ions on In e ac i e In elligen Sys ems (TiiS), 9(2-3), 1-30.
Saxena, A., P asad, M., Gup a, A., Bha ill, N., Pa el, O. P., Tiwa i, A., . . . Lin, C.-T. (2017). A
e iew o clus e ing echniques and de elopmen s. Neu ocompu ing, 267, 664-681.
Sca idi Abba e, C., Di Nuo o, S., & Ga o, M. (2017). A COMPARISON BETWEEN
MOTIVATIONS AND PERSONALITY TRAITS IN RELIGIOUS TOURISTS AND CRUISE
SHIP TOURISTS.
Scha e , J. B., F ankowski, D., He locke , J., & Sen, S. (2007). Collabo a i e il e ing
ecommende sys ems. In The adap i e web: me hods and s a egies o web
pe sonaliza ion (pp. 291-324): Sp inge .
Scha e , J. B., Kons an, J., & Riedl, J. (1999). Recommende sys ems in e-comme ce. Pape
p esen ed a he P oceedings o he 1s ACM con e ence on Elec onic comme ce.
Schein, A. I., Popescul, A., Unga , L. H., & Pennock, D. M. (2002). Me hods and me ics o cold-
s a ecommenda ions. Pape p esen ed a he P oceedings o he 25 h annual
in e na ional ACM SIGIR con e ence on Resea ch and de elopmen in in o ma ion
e ie al.
Sche e , K. R. (1979). Pe sonali y ma ke s in speech: Camb idge Uni e si y P ess.
Schmid -Belz, B., Nick, A., Poslad, S., & Zip , A. (2002). Pe sonalized and loca ion-based mobile
ou ism se ices.
Schneide , P. P., & Vog , C. A. (2012). Applying he 3M model o pe sonali y and mo i a ion o
ad en u e a ele s. Jou nal o T a el Resea ch, 51(6), 704-716.
Sco , D., & Lemieux, C. (2009). Wea he and clima e in o ma ion o ou ism. Commissioned
Whi e Pape o he Wo ld Clima e Con e ence 3. Wo ld Me eo ological O ganiza ion:
Gene a and Uni ed Na ions Wo ld Tou ism O ganiza ion: Mad id.
Sebas ia, L., Ga cia, I., Onaindia, E., & Guzman, C. (2009). e-Tou ism: a ou is
ecommenda ion and planning applica ion. In e na ional Jou nal on A i icial
In elligence Tools, 18(05), 717-738.
Selmi, A., B ahmi, Z., & Gammoudi, M. (2014). Mul i-agen ecommende sys em: S a e o he
a . Pape p esen ed a he P oc. o he 16 h in e na ional con e ence on in o ma ion
and communica ions secu i y.
Seno , C., Kos adino , D., Bouzid, M., Picaul , J., Aghasa yan, A., & Be nie , C. (2010). Analysis
o s a egies o building g oup p o iles. Pape p esen ed a he In e na ional
Con e ence on Use Modeling, Adap a ion, and Pe sonaliza ion.
Se kan, M., Neidha d , J., & We hne , H. (2019). Wha is he “Pe sonali y” o a ou ism
des ina ion? In o ma ion Technology & Tou ism, 21(1), 105-133.
Se kan, M., Neidha d , J., & We hne , H. (2020). PicTouRe-A Pic u e-Based Tou ism
Recommende . Pape p esen ed a he Fou een h ACM Con e ence on Recommende
Sys ems.
Shams, S., Ande son, D., & Lei h, D. (2021). Clus e -based bandi s: Fas cold-s a o
ecommende sys em new use s. Pape p esen ed a he P oceedings o he 44 h
In e na ional ACM SIGIR Con e ence on Resea ch and De elopmen in In o ma ion
Re ie al.
Re e ences
300
Sha ma, D., Banwala, E., Elouaghzani, I., & Ka a ya, R. (2023). Analysis o Pe sonalized
Tou ism Recommende Sys ems (2516-2314). Re ie ed om
Shaw, M. E. (1932). A compa ison o indi iduals and small g oups in he a ional solu ion o
complex p oblems. The Ame ican Jou nal o Psychology, 44(3), 491-504.
Shi, J. (2012). Time Alloca ion and he Wea he .
Sho , M., Tilak, S., Kuzne co a, I., Ma ens, B., & Akinkuolie, B. (2023). Gami ica ion in
mobile-assis ed language lea ning: A sys ema ic e iew o Duolingo li e a u e om
public elease o 2012 o ea ly 2020. Compu e Assis ed Language Lea ning, 36(3),
517-554.
Siegle , R. S. (1992). The o he Al ed Bine . De elopmen al psychology, 28(2), 179.
Sil a, D. L. C. d. (2023). De elopmen o an On ology o Tou is A ac ions o Recommending
Poin s o In e es in a G oup Recommende Sys em o Tou ism. (Mas e ), Poli écnico
do Po o,
Sil a, J. A., Fa ia, E. R., Ba os, R. C., H uschka, E. R., Ca alho, A. C. d., & Gama, J. (2013). Da a
s eam clus e ing: A su ey. ACM Compu ing Su eys (CSUR), 46(1), 1-31.
Smi h, K. A., Dennis, M., Mas ho , J., & Tin a e , N. (2019). A me hodology o c ea ing and
alida ing psychological s o ies o con eying and measu ing psychological ai s.
Use Modeling and Use -Adap ed In e ac ion, 29, 573-618.
Smi h, S. L. (2014). Tou ism analysis: A handbook: Rou ledge.
Smi h, V. L. (2012). Hos s and gues s: The an h opology o ou ism: Uni e si y o Pennsyl ania
P ess.
Smy h, B. (2007). Case-based ecommenda ion. In The adap i e web: Me hods and s a egies
o web pe sonaliza ion (pp. 342-376): Sp inge .
Sojah ood, Z. B., Taleai, M., & Cheng, H. (2023). Hyb id POI G oup Recommende Sys em
based on g oup ype in LBSN. Expe sys ems wi h applica ions, 119681.
Song, X., Liu, Z., Meng, L., Chu, D., Yu, J., & Sheng, Q. Z. (2024). Accu a e POI Recommenda ion
o Random G oups Wi h Imp o ed G aph Neu al Ne wo ks and Mul i-nego ia ion
Model. Resea ch Squa e. Re ie ed om h ps://asse s-
eu. esea chsqua e.com/ iles/ s-5297395/ 1_co e ed_999c43eb-0ece-435a-b de-
6ea8b56 2 .pd ?c=1733113957
Sönmez, S. F., & G ae e, A. R. (1998a). De e mining u u e a el beha io om pas a el
expe ience and pe cep ions o isk and sa e y. Jou nal o T a el Resea ch, 37(2), 171-
177.
Sönmez, S. F., & G ae e, A. R. (1998b). In luence o e o ism isk on o eign ou ism
decisions. Annals o ou ism esea ch, 25(1), 112-144.
Sposi o, V., Hand, M., & Ska pness, B. (1983). On he e iciency o using he sample ku osis in
selec ing op imal lpes ima o s. Communica ions in S a is ics-simula ion and
Compu a ion, 12(3), 265-272.
Su, L., Cheng, J., & Huang, Y. (2021). How do g oup size and g oup amilia i y in luence ou is
sa is ac ion? The media ing Role o pe cei ed alue. Jou nal o T a el Resea ch, 60(8),
1821-1840.
Suhud, U. (2015). A s udy o examine he ole o en i onmen al mo i a ion and sensa ion
seeking pe sonali y o p edic beha io al in en ion in olun ee ou ism.
In e na ional Jou nal o Resea ch, 4(1), 17-29.
Susi, T., Johannesson, M., & Backlund, P. (2007). Se ious games: An o e iew. Skö de:
Ins i u ionen ö kommunika ion och in o ma ion, 1-28.
Tabak, G., Canik, S., & Gune en, E. (2021). A e you a aid o a el du ing COVID-19?
Uni e si y o Sou h Flo ida M3 Cen e Publishing, 5(2021), 25.
Tan, W.-K. (2020). Des ina ion selec ion: in luence o ou is s’ pe sonali y on pe cei ed
a el cons ain s. Jou nal o Vaca ion Ma ke ing, 26(4), 442-456.
Re e ences
301
Tan, W.-K., & Tang, C.-Y. (2013). Does pe sonali y p edic ou ism in o ma ion sea ch and
eedback beha iou ? Cu en issues in ou ism, 16(4), 388-406.
Teng, C.-I., & Chen, Y.-J. (2009, 2009). Pe sonali y Face s and Cus ome Loyal y in Online
Games. Pape p esen ed a he The 9 h In e na ional Con e ence on Elec onic
Business.
Thomas, K. W. (2008). Thomas-kilmann con lic mode. TKI P o ile and In e p e i e Repo ,
1(11).
Tin a e , N., & Mas ho , J. (2015). Explaining ecommenda ions: Design and e alua ion. In
Recommende sys ems handbook (pp. 353-382): Sp inge .
Tkalcic, M., & Chen, L. (2015). Pe sonali y and ecommende sys ems. In Recommende
sys ems handbook (pp. 715-739): Sp inge .
Tkalcic, M., & Chen, L. (2022). Pe sonali y and ecommende sys ems. Recommende Sys ems
Handbook; Ricci, F., Rokach, L., Shapi a, B., Eds, 757-787.
Tkalčič, M., De Ca olis, B., de Gemmis, M., Odić, A., & Koši , A. (2016). In oduc ion o
emo ions and pe sonali y in pe sonalized sys ems. In Emo ions and Pe sonali y in
Pe sonalized Se ices (pp. 3-11): Sp inge .
Tkalcic, M., Kuna e , M., Koši , A., & Tasic, J. (2011). Add essing he new use p oblem wi h a
pe sonali y based use simila i y measu e. Pape p esen ed a he Fi s In e na ional
Wo kshop on Decision Making and Recommenda ion Accep ance Issues in
Recommende Sys ems (DEMRA 2011).
Tkalcic, M., Kuna e , M., Tasic, J., & Koši , A. (2009, 2009). Pe sonali y based use simila i y
measu e o a collabo a i e ecommende sys em. Pape p esen ed a he P oceedings
o he 5 h Wo kshop on Emo ion in Human-Compu e In e ac ion-Real wo ld
challenges.
Tondello, G. F., O ji, R., & Nacke, L. E. (2017). Recommende sys ems o pe sonalized
gami ica ion. Pape p esen ed a he Adjunc Publica ion o he 25 h Con e ence on
Use Modeling, Adap a ion and Pe sonaliza ion.
To es, R., & Skillico n, P. (2004). Mon ezuma’s e enge: How sani a ion conce ns may inju e
Mexico’s ou is indus y. Co nell ho el and es au an adminis a ion qua e ly,
45(2), 132-144.
Tsionas, A., & Sa a zemi, M. (2023). Educa ional S o y-Based Game o Cap u ing he
Lea ne 's Pe sonali y. Pape p esen ed a he Eu opean Con e ence on Games Based
Lea ning.
Tunguz, B. (2018). Big Fi e Pe sonali y Tes . Re ie ed om
h ps://www.kaggle.com/da ase s/ unguz/big- i e-pe sonali y- es /da a
Tuzhilin, A., Ko en, Y., Benne , J., Elkan, C., & Lemi e, D. (2008). La ge-scale ecommende
sys ems and he ne lix p ize compe i ion. Pape p esen ed a he KDD P oceedings.
UNWTO. (2001). Thesau us on Tou ism and Leisu e Ac i i ies. In Thesau us on Tou ism and
Leisu e Ac i i ies: Uni ed Na ions Wo ld Tou ism O ganiza ion.
Van de Hoek, W., & Woold idge, M. (2008). Mul i-agen sys ems. Founda ions o A i icial
In elligence, 3, 887-928.
Van Lank eld, G., Sp onck, P., Van den He ik, J., & A n z, A. (2011). Games as pe sonali y
p o iling ools. Pape p esen ed a he 2011 IEEE Con e ence on Compu a ional
In elligence and Games (CIG'11).
Van Se en, M., Pok ae , S., & Koolwaaij, J. (2004). Con ex -awa e ecommenda ions in he
mobile ou is applica ion COMPASS. Pape p esen ed a he In e na ional Con e ence
on Adap i e Hype media and Adap i e Web-Based Sys ems.
Van Wyhe, J. (2002). The au ho i y o human na u e: he Schädelleh e o F anz Joseph Gall.
The B i ish jou nal o he his o y o science, 35(1), 17-42.
Re e ences
302
Va his, N., Kons an opoulos, C., Pan ziou, G., & Ga alas, D. (2023). The Vaca ion Planning
P oblem: A mul i-le el clus e ing-based me aheu is ic app oach. Compu e s &
Ope a ions Resea ch, 150, 106083.
Ven u ini, E., Fa alla, V., & Innocen i, A. (2013). The ALBO P ojec : Vi ual Wo king
En i onmen s o he De ec ion o O ganiza ional Well-Being. In Riding he Hype
Cycle: The Resu gence o Vi ual Wo lds (pp. 13-25): B ill.
Vigolo, V. (2017). Olde Tou is s’ T a el Planning Beha io . In Olde Tou is Beha io and
Ma ke ing Tools (pp. 63-84): Sp inge .
Villamiza , M., Ga cés, O., Cas o, H., Ve ano, M., Salamanca, L., Casallas, R., & Gil, S. (2015).
E alua ing he monoli hic and he mic ose ice a chi ec u e pa e n o deploy web
applica ions in he cloud. Pape p esen ed a he 2015 10 h Compu ing Colombian
Con e ence (10CCC).
Vincia elli, A., & Mohammadi, G. (2014). A su ey o pe sonali y compu ing. IEEE
T ansac ions on A ec i e Compu ing, 5(3), 273-291.
Vinu ha, H., Poo nima, B., & Saga , B. (2018). De ec ion o ou lie s using in e qua ile ange
echnique om in usion da ase . Pape p esen ed a he In o ma ion and decision
sciences: P oceedings o he 6 h in e na ional con e ence on ic a.
Volo, S. (2021). The expe ience o emo ion: Di ec ions o ou ism design. Annals o ou ism
esea ch, 86, 103097.
W. Collie , R., O'Neill, E., Lillis, D., & O'Ha e, G. (2019). MAMS: Mul i-Agen Mic oSe ices.
Pape p esen ed a he Companion P oceedings o The 2019 Wo ld Wide Web
Con e ence.
Wallinge , S., G undne , L., Majic, I., & Lampol shamme , T. J. (2023). Agen -Based Modelling
o Sus ainable Tou ism. Pape p esen ed a he In o ma ion and Communica ion
Technologies in Tou ism 2023: P oceedings o he ENTER 2023 eTou ism
Con e ence, Janua y 18-20, 2023.
Wal on, D., & K abbe, E. C. (1995). Commi men in dialogue: Basic concep s o in e pe sonal
easoning: SUNY p ess.
Wang, A. I., & Tahi , R. (2020). The e ec o using Kahoo ! o lea ning–A li e a u e e iew.
Compu e s & Educa ion, 149, 103818.
Wang, Z., Yu, Y., Zheng, W., Ma, W., & Zhang, M. (2024). MACRec: A Mul i-Agen Collabo a ion
F amewo k o Recommenda ion. Pape p esen ed a he P oceedings o he 47 h
In e na ional ACM SIGIR Con e ence on Resea ch and De elopmen in In o ma ion
Re ie al.
Wa is, M., Fakha , M. Z., Gulsoy, M., Yalcin, E., & Bilge, A. (2024). A No el P e-P ocessing
Technique o Comba Popula i y Bias in Pe sonali y-Awa e Recommende Sys ems.
IEEE Access.
Wecke , A. J., Ku lik, T., & S ock, O. (2016). Dynamic pe sonaliza ion based on mobile beha io :
om pe sonali y o pe sonaliza ion: a bluep in . Pape p esen ed a he P oceedings
o he 18 h In e na ional Con e ence on Human-Compu e In e ac ion wi h Mobile
De ices and Se ices Adjunc .
Weinbe g, R. S., & Gould, D. (2023). Founda ions o spo and exe cise psychology: Human
kine ics.
Wickens, E. (2002). The sac ed and he p o ane: A ou is ypology. Annals o ou ism
esea ch, 29(3), 834-851.
Wi en, I. H., F ank, E., Hall, M. A., & Pal, C. (2016). Da a Mining: P ac ical Machine Lea ning
Tools and Techniques: Else ie Science.
Woold idge, M., & Jennings, N. R. (1995). In elligen agen s: Theo y and p ac ice. The
knowledge enginee ing e iew, 10(2), 115-152.

[Document text truncated for crawler view.]