ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
A Me hodology o S uc u ed On ology Cons uc ion
applied o In elligen T anspo a ion Sys ems
D. G ego a, S. To alb, T. A izac, F. Ba e ob, R. G ego d, J. Rodasd,
M. A zamendiaa
aLabo a o y o Dis ibu ed Sys ems, Facul y o Enginee ing, Na ional Uni e si y o
Asuncion, 2060 Isla Bogado, Luque, Pa aguay
bDepa men o Elec onics Enginee ing, Uni e si y o Se ille, 41092 Se ille, Spain
cDepa men o Telema ic Enginee ing, Uni e si y o Se ille, 41092 Se ille, Spain
dLabo a o y o Powe and Con ol Sys ems, Facul y o Enginee ing, Na ional Uni e si y
o Asuncion, 2060 Isla Bogado, Luque, Pa aguay
Abs ac
The numbe o compu e s ins alled in u ban and anspo ne wo ks has
g own emendously in ecen yea s, also he local p ocessing capabili ies and
digi al ne wo king cu en ly a ailable. Howe e , he he e ogenei y o exis ing
equipmen in he ield o ITS (In elligen T anspo a ion Sys ems) and he
la ge olume o in o ma ion hey handle, g ea ly hinde he in e ope abili y
o he equipmen and he design o coope a i e applica ions be ween de ices
cu en ly ins alled in u ban ne wo ks. While he dynamic disco e y o in-
o ma ion, composi ion and in oca ion o se ices h ough in elligen agen s
a e a po en ial solu ion o hese p oblems, all hese echnologies equi e in-
elligen managemen o in o ma ion lows. In pa icula , i is necessa y o
wean hese in o ma ion lows o he echnologies used, enabling uni e sal in-
e ope abili y be ween compu e s, ega dless o he con ex in which hey
a e loca ed. The main objec i e o his pape is o p opose a sys ema ic
me hodology o c ea e on ologies, using me hods such as a seman ic clus-
e ing algo i hms o e ie al and ep esen a ion o in o ma ion. Using he
p oposed me hodology, an on ology will be de eloped in he ITS domain.
This on ology will se e as he basis o seman ic in o ma ion o a SS (Se-
man ic Se ice) ha allows he connec ion o new equipmen o an u ban
∗Co esponding au ho
Email add ess: [email p o ec ed] (D. G ego )
P ep in submi ed o Compu e S anda s & In e aces Oc obe 14, 2015
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
ne wo k. The SS uses he CORBA s anda d as dis ibu ed communica ion
a chi ec u e.
Keywo ds: In elligen T anspo a ion Sys ems, On ology, Clus e ing,
In o ma ion Re ie al, Collabo a ion, CORBA, Dis ibu ed Sys ems,
S a is ical Da a Analysis.
1. In oduc ion
The eal- ime es ima ion o a ic pa ame e s and he con ol ope a ions
cons i u e a challenge o con ol o u ban a ic sys ems (Chen and Cheng,
2010). Un il now, he equipmen s ins alled in u ban ne wo ks usually wo k in
a cen alized way, p o iding in o ma ion o he a ic con ol cen e h ough
he u ban da a ne wo k and pe o ming ac ions acco ding o he decisions o
an ope a o a he con ol cen e . Howe e , he enhancemen s o anspo
equipmen s due o he e olu ion o elec onics and da a ne wo ks allow hem
o sha e in o ma ion and wo k coope a i ely. The main challenge in he de-
sign and ope a ion o ITS is in o ma ion exchange, which is a di icul ask
in highly dis ibu ed sys ems. F om a echnical s andpoin , he e a e di i-
cul ies in in eg a ing in o ma ion using complian s anda ds and connec ing
mul iple sys ems, especially when conside ing he complexi y and olume o
in o ma ion lows in ol ed in he ield o ITS, whe e bo h, he ha dwa e as
he da a gene a ed a e highly he e ogeneous (To al e al.,2010). The e o e
i is necessa y o op imize he in e ope abili y, secu i y and e iciency o p o-
cesses and de ices which a e pa o he ITS by de eloping new echnologies.
Mo e speci ically, i is necessa y o analyze he needs o he anspo and
logis ics om a mul imodal pe spec i e, and o design new sys ems and ools
able o p o ide “highe in elligence” in he p ocess o in o ma ion exchange
and in e ope abili y be ween de ices. Dis ibu ed sys ems a e well known o
hei di icul y o in e ope a ion among agen s, which jus i ies he in e es in
uni ied so wa e pla o ms (Wang e al.,2006). SOA (Se ice-O ien ed A chi-
ec u e) is p esen ed as an a ac i e al e na i e o enable in e ope abili y o
sys ems and he euse o esou ces. Bu SOA applica ions ace many secu i y
p oblems du ing design and de elopmen (Qu e al.,2010). In SOA a chi-
ec u es, he WS (Web Se ices) a e a commonly used echnology. WS use
SOAP (Simple Objec Access P o ocol) as he communica ion p o ocol be-
ween a ious se ices. SOAP is an XML-based p o ocol. Howe e , p ocess-
ing la ge SOAP messages signi ican ly educes sys em pe o mance, causing
2
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
bo lenecks in compa ison wi h o he echnologies like CORBA (Tekli e al.,
2012). This ep esen s a p oblem in wi eless communica ion ne wo ks (Phan
e al.,2008) and in he ITS ield, whe e he numbe o connec ed de ices
is g owing o e ime. In p ac ice, SOA-based applica ions a e no always
success ul as mos o hem a e done on an ad-hoc basis, and p ima ily based
on pe sonal expe iences (Guo e al.,2010). Al hough companies a e inc eas-
ing hei dependence owa ds SOA, hese sys ems a e s ill in an imma u e
ea ly s age wi h impo an secu i y p oblems (Kabbani e al.,2010). The
common p oblem in all he men ioned echnologies is he in e ope abili y
be ween se ices and de ices ha a e pa o ITS, due o he di e ences in
he in o ma ion ep esen a ion and seman ics. The use o on ologies in his
ield would be a solu ion o enable euse o domain knowledge and o gene a e
sma clien s. Agen s ha sha e seman ic in o ma ion could use his on olog-
ical in o ma ion o espond o eques s DD (De ice-De ice), se e as inpu
o o he se ices, enable euse o domain knowledge o wo k coope a i ely
wi h o he exis ing on ologies.
A me hodology o de ine an on ology in he ield o ITS is p oposed. The
on ology will be used in a CORBA-complian Seman ic Se ice, which allows
inding se ices in a dis ibu ed en i onmen . The de eloped on ology will
se e as ini ial Da aBase o he in elligen sys em o seman ic managemen ,
whe e he ha dwa e de ices can exchange in o ma ion h ough a commu-
nica ion sys em and wo k coope a i ely. Sec ion 2p o ides an o e iew o
p e ious ela ed wo k. In Sec ion 3, he p oposed me hodology is in oduced,
using as a s a ing poin Sys ema ic Li e a u e Re iew (SLR) echniques, and
hen applying seman ic analysis echniques and s a is ical da a analysis o
build he ITS on ology. Sec ion 4de ails he ob ained esul s wi h he p o-
posed me hodology, es ing he esul ing on ology in a CORBA dis ibu ed
en i onmen . Finally, Sec ion 5shows he conclusions and u u e wo k.
2. Rela ed Wo k
One o he main challenges in he ITS is he coope a i e a ic. The
idea o coope a ion wi hin ITS was ini ia ed by he concep o coope a i e in
au oma ed highways whe e ehicles ecei e inpu signals om he oad en i-
onmen . The i s ideas documen ed on au oma ed highways we e p esen ed
in 1960 by he esea ch labo a o y o he Gene al Mo o s (Ga dels,1960). A
coope a i e a ic sys em makes use o da a as soon as hey a e collec ed,
au oma ing decision making in si ua ions ha equi e he in elligen in e -
3
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
en ion o ITS en i onmen . (Soa es e al.,2009) p esen a s a egy in da a
dissemina ion o coope a i e sys ems, de ending ha di usion policies plays
a de e mining ole in he sp ead o ITS o he e icien in o ma ion p opa-
ga ion. Indeed, he main objec i e o he coope a i e d i ing is o ocus on
p e en ion and ea ly de ec ion o isks. Howe e , his s udy does no speci y
how o ind o main ain in o ma ion. (Rockl and Robe son,2010) a gue
ha he success o coope a i e ITS applica ions is mainly a ec ed by he ex-
change o in o ma ion be ween dis ibu ed nodes. Acco ding o au ho s, he
ansmission o la ge amoun o in o ma ion con as s o he limi ed band-
wid h o he channels ha end o be sha ed by all nodes pa icipa ing in
he ITS. Bu he ex ac ion and in e p e a ion o he in o ma ion is ou o
he scope o he s udy. The e o e, i is necessa y o de elop e icien he -
e ogeneous al e na i es o inc ease he e ec i e capaci y o he ITS and o
imp o e he e iciency o he anspo sys ems. The solu ion lies mainly
in he coope a i e commi men o selec ele an pieces o in o ma ion o
dissemina ion acco ding o hei alue.
Wi h he inc easing de elopmen o elec onics and he possibili y o using
embedded sys ems wi h inc easing p ocessing capabili ies, he concep o co-
ope a ion has been ex ended om he o iginal idea o coope a i e d i ing
o he cu en ITS dis ibu ed sys ems. The main idea o coope a ion in
ITS dis ibu ed sys ems is based on he collabo a ion o ehicle d i ing wi h
a ailable se ices in u ban, subu ban, me opoli an and u al a eas, whe e
ehicles in e ac wi h he en i onmen , and he en i onmen i sel ac s in el-
ligen ly based on a ic e en s. (Mi opoulos e al.,2010) p esen ed a sys em
called WILLWARN (Wi eless Local Dange Wa ning) based on ecen and
u u e ends in coope a i e d i ing allowing elec onic secu i y o p e en
isks h ough “Vehicle-Haza d” de ec ion applica ions on-boa d, V2V (Vehi-
cle o Vehicle) and V2I (Vehicle o In as uc u e) communica ions. One o
he main causes o oad acciden s is he excessi e and slow eac ion o he
d i e in c i ical si ua ions. Howe e , he sys em p oposed by Mi opoulos is
exclusi ely ocused on managing messages ale ing he d i e o he dange
in ad-hoc basis, igno ing he quali y and p esen a ion o in o ma ion.
(Thomas and an Be kum,2009) p oposed a p edic ion scheme o e-
cu ing a ic e en s based on da a collec ed a u ban in e sec ions. They
a gue ha i is necessa y he managemen o e en s on demand in case o
possible inciden s, bu hey do no alida e he esul s o he analysis wi h
eal da a o inciden de ec ion, and hey do no de ine how he in o ma ion
is collec ed, shown o s o ed. The main challenges in cu en ITS dis ibu ed
4
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
a chi ec u es, whe e in o ma ion plays an impo an ole, a e he he e ogene-
i y o so wa e, ha dwa e de ices, and communica ion ne wo ks. In he case
o ha dwa e de ices i is usual he incompa ibili y in he da a ep esen a ion,
he p oblems o synch oniza ion and he wide a ie y o con olle s and p o-
cesso s. So wa e applica ions and se ices ha e p oblems caused by he exis-
ence o mul iple p og amming languages, di e en e sions o he same appli-
ca ion o se ice, he compe i ion be ween p op ie a y and ee/open sou ce
so wa e as well as p oblems o unde s anding and dis ibu ed Da aBases
complexi y. Finally, he e ogenei y in communica ion ne wo ks is mainly due
o he wide a ie y o ne wo k p o ocols, and he deploymen o dis ibu ed
ne wo ks, in some cases incompa ible wi h adi ional ne wo ks.
To o e come hese d awbacks, on ologies can be an impo an issue in he
u u e o ITS. One o he main ad an ages o he in eg a ion o on ologies
in ITS is he in elligen and secu e seman ic loca ion o se ices wi h ce ain
cha ac e is ics and p ope ies. F om he poin o iew o in e ope abili y
be ween de ices om di e en endo s and pla o ms, he mos s iking ad-
an age is he in elligen in o ma ion e ie al. Se ices can be published in
desc ip i e on ologies and de ices can make use o da a and me ada a om
di e en kinds o un ime a ic e en s. While mo e s uc u ed is he se ices
in o ma ion, mo e accu a e, as and sma hey can be ound. Me ada a can
p o ide some seman ics o his p oblem since on ologies p o ide a concep-
ual amewo k o exploi h ough me ada a exchange schemes. Nume ous
p e ious s udies ha e made use o me ada a o imp o e implemen a ion o
collabo a i e applica ions in di e en scena ios. (Ga c´ıa e al.,2012) p esen
a con ex model based on an on ology which akes a combined app oach o
model he con ex in o ma ion used by anspo se ices.
The modeled dis ibu ed in o ma ion is ela ed o a p ima y con ex
abou he loca ion, ime, iden i y and quali y o se ices, bu applied only
o a se ice o loca ion o pa king spaces. Thanks o he p oposed scena io,
hey demons a e ha con ex in o ma ion gene a ed om au onomous dis-
ibu ed sou ces can be ep esen ed using a common da a model and can
be s uc u ed acco ding o a common on ology. The esul ing da a can be
sha ed, associa ed, used, o easoned. (Chen e al.,2008) p oposed he
design and implemen a ion o a amewo k o public anspo . They in-
clude a mechanism o da a collec ion h ough WS which a e speci ically used
o planning ou es. Howe e , he use o WS usually based on SOAP and
XML may cause excessi e bandwid h consump ion o mo e complex sys ems
whe e he e is a big demand o se ices. (Fe nandez and Ossowski,2011)
5

ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
suppo he assump ion ha he use o MAS (Mul i Agen Sys ems) enables
a decoupled design and he implemen a ion o di e en modules (agen s),
encou aging euse o simila on ological domains, educing he de elopmen
e o and inc easing sys em eliabili y ( euse o exis ing se ices). They o-
cus he s udy on a se ice o ien ed mul i-agen a chi ec u e o cons uc ing
ad anced DSSs (Decision Suppo Sys ems) in anspo managemen . How-
e e , hey do no speci y how o use he in o ma ion as a ool o how o
wo k coope a i ely wi h o he exis ing on ologies. (Te ziyan e al.,2010) de-
ail he equi emen s and he necessa y a chi ec u e o a ic managemen
sys ems, showing how such a sys em can be bene icial om he seman ic
poin o iew h ough echnologic agen s bu ques ioning how his sys em
can be combined wi h da a p ocessing and au oma ed ools. A sys em o
in o ma ion e ie al based on a uzzy on ological amewo k was p oposed
by (Zhai e al.,2008). The p oposed amewo k is composed o h ee ele-
men s: concep s, p ope ies o concep s and alues o p ope ies, being he
p ope y alue any s anda d da a ype o linguis ic alues o uzzy concep s.
The main d awback o his amewo k is ha he in o ma ion e ie al sys-
em is p ima ily ocused on in o ma ion abou a ic acciden s, lea ing aside
o he key issues such as in e ope abili y be ween de ices o he e ogenei y
o in o ma ion. A coope a i e a ic sys em should be able o sol e com-
plex p oblems using en i onmen al da a and me ada a. The ITS equipmen s
should be p epa ed o lea n om he en i onmen and change hei cha ac-
e is ics based on e en s. Addi ionally, hey mus be able o in e ac wi h
each o he , o ming mul i-agen sys ems o achie e objec i es. In his pape
i is p oposed an in elligen solu ion in he eco e y and managemen o he -
e ogeneous in o ma ion in o de o build an on ology using a axonomy as
he s a ing poin o he s udy. The on ology will se e o o ganize and o e
a me ada a based se ice sp ead ac oss he a ic ne wo k.
One o he i s s eps in he on ology cons uc ion is undoub edly he IR
(In o ma ion Reco e y). Due o he la ge amoun o a ailable in o ma ion,
building on ologies om sc a ch and manually would equi e a lo o ime
and e o . The e o e, i is necessa y o inco po a e scien i ic echniques in
he analysis and dynamic selec ion o in o ma ion o p o ide a logical s uc-
u e. Scien i ic Sys ema ic Li e a u e Re iew (SLR) is he ield o s udy
ha ies o analyze and in eg a e essen ial in o ma ion o he p ima y e-
sea ch s udies on pa icula opic, in a pe spec i e o se uni a y syn hesis.
SLR has become an impo an esea ch me hodology o he eco e y and
collec ion o in o ma ion (Hall e al.,2012). The aim o SLR is he iden i i-
6
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
ca ion, e alua ion and in e p e a ion o all ele an esea ch s udies abou a
pa icula esea ch ques ion using igo ous me hods and speci ic algo i hms.
(Zhang e al.,2011) a gue ha he accu acy and p eciseness in he in o ma-
ion sea ch p ocess is ac ually a c i ical poin ha dis inguishes sys ema ic
e iews om he adi ional ad-hoc li e a u e e iews. They ha e de eloped
a sys ema ic app oach based on e idence o he de elopmen and imple-
men a ion o op imal sea ch s a egies on digi al li e a u e. The p oposed
app oach inco po a es he concep o “quasi-gold s anda d” (QGS), which
is he collec ion o known s udies, and he co esponding “quasi-sensi i i y”
in he sea ch p ocess o e alua e i s pe o mance. The e a e se e al wo ks
abou me hodologies o de eloping on ologies. (G uninge and Fox,1995),
p oposed a me hodology o design and e alua e an on ology ha i s in u-
i i ely iden i ies he possible applica ions whe e he on ology can be used.
They use a se o ques ions called “compe ency ques ions” o de e mine he
scope o he on ology and o ex ac key concep s, p ope ies, ela ions and
axioms. A mo e sys ema ic app oach o he cons uc ion o an on ology
om sc a ch is he so called Me hon ology (Fe nandez e al.,1997). This is
pe haps one o he mos comple e p oposed me hods and conside s he de el-
opmen o on ologies as a compu e p ojec . I includes ac i i ies o p ojec
planning, quali y esul s, documen a ion, e c., and allows he building o new
on ologies o eusing exis ing ones. (Chand asega an e al.,2013) applied a
o mal concep analysis me hodology o de elop a domain-speci ic on ology.
They used a o mal concep analysis o iden i y simila i ies among a ini e
se o objec s based on hei p ope ies, p o iding a concep ual hie a chical
clus e ing. Howe e , he abo e me hods lack o ools o IR and SLR. I is
necessa y o conside IR and SLR as pa o he me hodology o on ology
c ea ion in o de o a oid bias in he esul ing on ology, as he me hodology
p oposed in his pape .
3. METHODOLOGY
Fig. 1shows a block diag am o he p oposed me hodology o de eloping
he on ology. The main objec i e o he p oposed me hodology is o disco e
ITS se ices based on common pa e ns among he da a, wi h he inal aim
o ob aining class hie a chies in he “Building he On ology” block.
The p oposed me hodology includes se e al au oma ed me hods o de-
eloping me a-analysis echniques on documen s. The s a ing poin is a
axonomy ha summa izes he main opics o a domain ield and a collec-
7
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
Figu e 1: Block Diag am: P oposed Me hodology.
ion o documen s ep esen ing he majo esea ch ends in he ITS a ea.
Based on he p oposed me hods, a comple e on ology o se ices and se ice
con aine s in he domain o ITS can be buil . The esul s a e a concep ual
scheme ha can be exploi ed h ough me ada a exchange among de ices and
embedded applica ions in dis ibu ed u ban sys ems. The ollowing subsec-
ions desc ibe in de ail each block lis ed in he gene al scheme o he p oposed
me hodology.
3.1. Taxonomy De ini ion
The i s s ep o de eloping an on ology consis s o ob aining a se o
basic concep s o classes ha de ine a speci ic domain, he ITS ield in his
case. Typically, his s ep in ol es he sea ch o a se o keywo ds co e ing
all he opics and issues ela ed o he a ge domain. Howe e , in he case
o he ITS ield, se e al o ganiza ions like U.S. DOT (Uni ed S a es Depa -
men s o T anspo a ion) ha e p e iously explo ed his ield in de ail (RITA
U.S. DoT,2015). Mo e speci ically, he Resea ch and Inno a i e Technology
Adminis a ion (RITA) coo dina es he U.S. Depa men o T anspo a ion’s
esea ch p og ams and i is in cha ge o he ad ances in he deploymen
o c oss-cu ing echnologies o imp o e he anspo a ion sys em (RITA,
2015), (USDOT,2015). As pa o hei ac i i ies, hey ha e de eloped
a axonomy o he ITS ield conside ing se e al Le els o De ail (LoD), as
shown in Fig. 2, which ep esen s a pa o he RITA U.S. DoT axonomy.
In his s udy, i has been conside ed his axonomy un il he LoD 4, which
p o ides a collec ion o 77 con aine s o se ices. This le el o de ail has been
8
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
Figu e 2: Pa o RITA U.S. DoT axonomy (un il LoD 5).
chosen because i is an in e media e poin be ween p e ious oo gene ic and
subsequen oo de ailed con aine s o se ices.
3.2. Selec ion o he Collec ion o Documen s
The nex s ep is he selec ion o he ele an in o ma ion o apply seman-
ic echniques. The 10 jou nals wi h he highes Impac Fac o s (IF) in he
ield o he ITS and, o each one, he 30 mos impo an publica ions o
he las 10 yea s (2005 o 2015) has been collec ed, gi ing as a esul 300
publica ions, as shown in Table 1.
No ice ha he selec ed in o ma ion is g ouped in collec ions o 30 pa-
pe s. One o he d awbacks o using a collec ion o documen s is ha he
weigh o each keywo d in each pape is di e en . One possible solu ion o
o e come his issue is he IR eedback echnique o ele ance (Sal on and
Buckley,1990). The main idea in his echnique is ha once ce ain e-
ie ed documen s ha e been conside ed as ele an o i ele an by he use ,
he p o ided in o ma ion is used o adap he que y so mo e ele an docu-
men s a e e ie ed in a subsequen sea ch. Howe e , he p ocess o al e ing
a que y in he di ec ion o ele an documen s is an e ec i e echnique in
in o ma ion e ie al o an en i e documen , bu no o speci ic pa s o i .
This pape p oposes a no el me hod o he iden i ica ion o pa ag aphs in
he collec ion o documen s as an al e na i e o he basic uni o analysis.
3.3. Disc imina ion o Pa ag aphs wi h Keywo ds
The main objec i e o he p oposed Disc imina ion o Pa ag aphs wi h
Keywo ds (DPK) is o e ie e only he mos ele an in o ma ion in he
9
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
In his pape , he a e age-linkage algo i hm has been chosen because o i s
obus ici y (E e i e al.,2011), i s highe pe o mance (Li e al.,2009) and
he quali y o p o ided clus e s (Sileshi and Gamback,2009).
In he a e age-linkage, he dis ance be ween wo clus e s is de ined as
he a e age dis ance be ween pai s o obse a ions, one in each clus e .
The a e age-linkage commonly joins clus e s wi h small a ia ions and ends
sligh ly o p oduce clus e s wi h he same a iance. The hie a chical clus e -
ing esul s can be g aphically ep esen ed as a ee-diag am o a dend og am.
S ep 2. Applica ion o he UPGMA Me hod o Build he Ul ame ic T ee
and o Ex ac Pai s/T iple Wo ds
One o he compu a ional challenges o his s udy is o ob ain a da a
s uc u e o help in he inal ep esen a ion o an on ology. One solu ion
p oposed by (Gibas and Jambeck,2001) was he implemen a ion o phyloge-
ne ic ees. In compu e science, he e is a da a s uc u e ha possesses he
p ope ies o phylogene ic ees called ul ame ic ees.
The dis ance be ween wo a bi a ies e ices xand yo T,disT(x, y),
is he sum o he weigh s o he edges composing he pa h om x o y
(Bockenhaue and Bonga z,2007). Gi en a ma ix o axa (subjec s o ob-
jec s), wo simple me hods o building ul ame ic ees can be used. The
i s one is called Unweigh ed Pai G oup Me hod wi h A i hme ic Mean
(UPGMA) and he second is Weigh ed Pai G oup Me hod wi h A i hme ic
Mean (WPGMA). Bo h o hem a e agglome a i e hie a chical me hods
using a e age-linkage echnique. The UPGMA is widely used in bioin o -
ma ics o de elop axonomies wi h nume ical da a ob ained om a se o
axa (Sokal and Snea h,1963). This me hod cons uc s he bo om-up phy-
logene ic ee om he lea es (se o axa). In he UPGMA me hod, dis ances
a e calcula ed using an a i hme ic a e age depending on he numbe o el-
emen s in each clus e . Basically bo h me hods, UPGMA and WPGMA,
wo k in he same way. The only di e ence is he unc ion o dis ance used
in he las s ep. WPGMA makes use o he weigh ed a e age, which ensu es
ha each axon is equally pa icipa ing in he inal esul . Wi h he dis ance
unc ion used by WPGMA, each axon con ibu es equally o he inal esul .
UPGMA and WPGMA di e in he inal esul bu no in he ma hema ical
mechanism o achie e i . Fo he me hodologies p oposed in his pape , he
UPGMA me hod is used because i is simple , as e and ha e been widely
used in he li e a u e.
The axonomy p oposed by he U.S. DOT and RITA, Fig. 2, consis s o
16

ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
wo majo g oups, one o hem ocused on he In elligen In as uc u es and
he o he one on In elligen Vehicles. The o al sample consis ed o 34,738
pa ag aphs. In he case o In elligen In as uc u e, he disc imina ed sam-
ple using he DPK me hod was o 1,519, disca ding he es o pa ag aphs
because o hei low ele ance acco ding o he conside ed con aine s o se -
ices.
Fig. 5de ails he numbe o pa ag aphs associa ed o he di e en con-
aine s o se ices included unde he gene al g oup o In elligen In as uc-
u es. I can be no iced a clea end o esea ch on T a ic Con ol. These
esul s can be clea ly explained by he inc easing in es men o public au-
ho i ies in he imp o emen o Road In as uc u e and Secu i y. Fig. 6
de ails he same esul bu o he case o In elligen Vehicles. A o al o 843
pa ag aphs we e disc imina ed om he ini ial sample using he p oposed
ools. Ob ained esea ch ends a e mo e balanced among he di e en con-
aine s o se ices, bu wi h mo e emphasis on Rou e Guidance. This is also
a expec ed esul since ou e guidance ools ha e become an impo an way
o alle ia ing conges ion in u ban anspo ne wo k and hey a e closely
ela ed o T a ic Con ol in he In elligen In as uc u es.
Table 3de ails he size educ ion a e applying DPK me hods measu ed
in pa ag aphs and MB. The educ ion o IIwas 95.63% while he educ ion
in In elligen Vehicles was 97.6%. A e applying IRWDP me hod and he
dimensionali y educ ion o LSA, i is possible o loca e keywo ds o each

Figu e 5: Rela i e F equency o Pa ag aphs abou In elligen In as uc u es.
17
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT

Figu e 6: Rela i e F equency o Pa ag aphs on In elligen Vehicles.
con aine o se ices in a ca esian coo dina e space. Table 4shows a educ-
ion o h ee dimensions o he pa icula case o “Su eillance” con aine
using he p edic i e analy ics ool, RapidMine 5 (Rapid-I,2012), (Lessmann
e al.,2008).
The main p oblem o a h ee dimensional ep esen a ion, is ha esul s
a e mo e di icul o be in e p e ed. Fo his eason, i is p e e able o con-
side only wo dimensions and assume he loose o in o ma ion in o de o
Table 3: Resul s a e applying DPK me hod.
Sample size educ ion a e DPK
ITS TOTAL PARA-
GRAPHS
FILTERED PARA-
GRAPHS
SAMPLE SIZE RE-
DUCTION
II 34 738 1 519 95.63%
IV 34 738 843 97.60%
Use ul Da a Size a e DPK
SIZE IN MB REDUCED RATE USEFUL DATA SIZE
II 13.2 95.63% 0.58 MB
IV 13.2 97.60% 0.58 MB
II: In elligen T anspo a ions, IV: In elligen Vehicles.
18
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
Table 4: LSA 3D - Da a Dimensionali y Reduc ion.
ExampleSe (150 examples, 1 special a ibu e, 3 egula a ibu es)
Row No. Wo d s d 1 s d 2 s d 3
1 T a ic 0.074 0.001 -0.042
2 Su eillance 0.077 0.037 0.009
3 Time 0.089 0.003 0.056
4 Da a 0.093 0.025 0.026
5 Sys em 0.077 0.079 0.021
6 Vehicle 0.073 -0.009 -0.066
7 Real 0.051 0.071 -0.031
.
.
..
.
..
.
..
.
..
.
.
150 wo d 150 · · · · · · · · ·
bene i he in e p e abili y o esul s. Table 5and Fig. 7de ail he dimen-
sionali y educ ion and he g aphical ep esen a ion conside ing only wo
dimensions o he same pa icula case o “Su eillance” con aine . In his
g aph, he diame e o each bubble ep esen s he simila i y be ween a ge
wo ds and colo o hese, ep esen each o he 150 e ms. Then, he a e age-
linkage model o agglome a i e hie a chical clus e ing is applied o ep esen
keywo ds o each con aine as a dend og am o ul ame ic ee using he
UPGMA me hod. This way da a is o ganized in o subca ego ies ha will be
in u n di ided in o he s un il eaching he desi ed le el o de ail. The ul-
ame ic dis ances a e hen hose ha mee he c i e ia o h ee poin s ( he
h ee-poin condi ion) (Deonie e al.,2005) which say: dis an ul ame ic
ee in Q, i he elemen s in each h ee-elemen -subse o Qcan be labeled
by x,y,zsuch ha :
d(x, y)≤d(x, z) = d(y, z).(2)
Acco ding o he ul ame ic ees, pai s and iple o wo ds ha e been
ex ac ed o build he on ology. Using pai s and iples o nea es wo ds, i is
possible o ex ac he in o ma ion ha la e i is used o build he on ology.
Fig. 9shows he pa icula esul o he case o “Su eillance” con aine .
Following he same p ocedu e wi h he es o con aine s ex ac ed om
LoD4 axonomy, he whole on ology is comple ed. Due o space limi a ions,
i is no possible o include he comple e ITS on ology o he ee diag ams.
19
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
Table 5: LSA 2D - Da a Dimensionali y Reduc ion.
ExampleSe (150 examples, 1 special a ibu e, 2 egula a ibu es)
Row No. Wo d s d 1 s d 2
1 T a ic 0.074 0.001
2 Su eillance 0.077 0.037
3 Time 0.089 0.003
4 Da a 0.093 0.025
5 Sys em 0.077 0.079
6 Vehicle 0.073 -0.009
7 Real 0.051 0.071
.
.
..
.
..
.
..
.
.
150 wo d 150 · · · · · ·
0.009 0.013 0.018 0.025 0.035 0.050 0.071 0.100 0.141
-0.10
-0.07
-0.04
-0.01
0.02
0.05
0.08
0.11
0.14
0.17
0.20
0.23
0.26
s d_1
s d_2
Figu e 7: Plo Sca e 2D - Su eillance Con aine .
Using he open sou ce on ology edi o P o ´eg´e (P o ´eg´e,2015), he de eloped
ITS on ology can be modeled in OWL o ma o po ed o o he s such as
RDF, RDFS, e c.
4. RESULTS
In his sec ion he esul s ha e been di ided in wo subsec ions. In he
i s one, i is a alida ion o he on ology buil . In he second, i is con-
duc ed se e al expe imen s o e alua e he pe o mance and scalabili y o
20
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
2UZ
Figu e 8: Con aine s o se ices in In elligen In as uc u es.
he low o in o ma ion on embedded sys ems ypically used in eal u ban
and dis ibu ed en i onmen s.
4.1. On ology alida ion
The axonomy de ined by he U.S. DOT and RITA add esses he classi-
ica ion o ITS applica ions. They p o ide a sys ema ic o ganiza ion o he
Figu e 9: Pa o ITS On ology - Su eillance Con aine .
21

ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
Figu e 10: Homonymy in he RITA Taxonomy.
ITS ield, gi ing names o g oups o elemen s and inal applica ions. A hie -
a chical s uc u al model connec s all e ms in he axonomy. Basically, his
axonomy conside s wo big ca ego ies: “In elligen In as uc u es” wi h 14
applica ions and “In elligen Vehicles” wi h 3 applica ions. Each o hese 17
applica ions is di ided in o sub-applica ions wi h a b ie summa y o hei
bene i s and in o ma ion ela ed o he a ea o in e es . Howe e , he axon-
omy is only a simple classi ica ion ha o e s he cos s and bene i s o each
applica ion, wi hou any seman ic o logical s uc u e in he da a exchange.
As a di e ence, he de eloped on ology “ITS. d s” adds a desc ip i e
logic. The da a and me ada a a e s o ed in eposi o ies, which p o ide access
o all in o ma ion on ITS applica ions and se ices disco e ed. The on ology
is able o cope wi h he p oblems ha RITA axonomy canno sol e, such
homonymy. Fig. 10 shows an example o homonymy p oblem in he case o
Su eillance, bo h sub-classes o A e ial Managemen and F eeway Manage-
men wi hin he ca ego y o In elligen In as uc u e. Any sys em seeking
a T a ic se ice abou Su eillance wi hin he axonomy would ecei e bo h
se ices, because i would be unable o dis inguish one o hem. Al hough
he axonomy con ibu es o he seman ics o a e m in he ocabula y, hey
do no de ine a ibu es be ween concep s and hus may cause con usion and
con lic s. As a di e ence, he ITS. d s on ology is iche in e ms o ela ions
be ween e ms. These ela ions allow o exp ess he in o ma ion wi hin he
domain wi hou he need o duplica ing e ms; a oiding homonyms.
Fig. 11 shows ha in he p oposed on ology, T a ic and In as uc u e
can be se ice con aine s, applica ions o se ices, belonging o he Su eil-
lance class. Simila ly, Su eillance is a sub-class o A e ial Managemen
22
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
Figu e 11: Homonymy solu ion in he ITS. d s on ology.
as well as F eeway Managemen and hese a e hemsel es sub-classes o
In elligen In as uc u e. As a di e ence o he axonomy case, he e he e
is no homonymy because hey ha e di e en meanings and he nodes a e
in di e en seman ic spaces. Thanks o namespaces, i is possible o a oid
ambigui ies in he esul . Nex igu es compa e he quan i y (Q y.) o con-
aine s/se ices p oposed by he U.S. DOT and RITA wi h he one ob ained
by he de eloped ITS. d s on ology, o In elligen In as uc u es, Fig. 8
and o In elligen Vehicles, Fig. 12. The quan i y o se ices disco e ed
abou In elligen In as uc u es in he de eloped on ology is 866, while he
RITA axonomy o e s a maximum o 144 se ices/applica ions. In he case
o In elligen Vehicles, he o al amoun o disco e ed se ices is 449, agains
he 6 se ices/applica ions o e ed by he RITA axonomy. The disco e ed
se ices may be used as he basis o de eloping new applica ions/se ices
in he ield o ITS. The on ology de eloped will se e as a e e ence ool o
in o ma ion acquisi ion and cons uc ion o knowledge base sys ems ha p o-
ide consis ency, eliabili y and accu acy when e ie ing in o ma ion. The
ITS. d s on ology enable sha ing he knowledge and enable he collabo a i e
wo k o unc ion as common medium o knowledge be ween di e en ac o s
in ol ed in a u ban, me opoli an and u al in as uc u e.
4.2. On ology Implemen a ion
The c ea ed on ology ITS. d s wi h he p oposed me hods is used as a
desc ip i e seman ic se ice con aine , and he low o in o ma ion is ea ed
as iple s SPO (Subjec , P edica e, Objec ) o an Seman ic Se ice (G e-
go e al.,2012) (Seman ic Communica ion Se ice on ology-based) capable
o managing he low o clien /se e eques s in dis ibu ed u ban en i on-
men s. An impo an measu e o check he pe o mance is he h oughpu
me hod as ollows:
Tpu kB =size(kB)
RTT(sec.),(3)
23
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
whe e RTT is he “Round T ip Time” in seconds. Howe e , his measu e is
no use ul because he on ological da a a e exp essed in iple s. Thus, he
p e ious me ic can be ex ended as ollows:
Tpu kT =no. iples/1000
RTT(sec.),(4)
which ep esen s he o al calcula ion on kiloT iple s o he on ology, o e
he RTT in seconds. To check, he o e all pe o mance has been es ed
s o ing he ob ained on ology in a Be keley Da aBase (O acle,2014) on a
PC-AMD A hlon (TM) 1200 MHz. The Seman ic Se ice is capable o p o-
iding he communica ion suppo on dis ibu ed en i onmen s in conjunc-
ion wi h a se o base lib a ies like Redland (D.,2011c) (RDF Language
Bindings) o in e ac wi h on ologies w i en in RDF and RDFS o ma s.
A Rap o pa se (D.,2011a) (RDF Syn ax Lib a y) is used o analyze he
sequences o symbols, de e mine he g amma ical s uc u e and as a que y
language, Rasqal (D.,2011b) and (RDF Que y Lib a y) o build and un
que ies. Bo h, Rasqal and Rap o a e designed o wo k wi h he Redland
lib a y. The goal o he dis ibu ed communica ion echnology used in hese
es s is o manage he on ological in o ma ion and in e ope a e wi h se ices
2UZ
Figu e 12: Con aine s o se ices in In elligen Vehicles.
24
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
Table 6: Pe o mance o he expe imen s.
Pe o mance o he expe imen 1
Pa sing and S o ing he ITS. d s scheme in Be keley Da aBase
Tpu kT A e age 3.57 kT/sec.
To al Time 806 ms.
To al Da a Size 625.58 kBy es.
Pe o mance o he expe imen 2
Read and analyze he empo al Model ecei ed om he Se e
T ans e Ra e (Tpu kB) 176.40 kB/sec.
To al Delay 16 ms.
To al Da a Size 2.84 kBy es.
Pe o mance o he expe imen 3
Delay esol ing he Que y and building he e-
sponse o sen o Clien
To al Delay 26 ms.
To al Da a Size o Send 1.05 kBy es.
To al Delay (Clien -Side) 41 ms.
disco e ed by he p oposed me hodologies in he p e ious sec ions. Fi s , i
is measu ed he pe o mance pa sing and s o ing he main ITS. d s scheme
in a Da aBase hos ed on he PC-AMD. The pe o mance was qui e s able
du ing he expe imen , Table 6.
In u ban en i onmen s, he di e en a ic se ices a e mos ly imple-
men ed in embedded de ices. The nex s ep was o es ima e he sys em pe -
o mance by adding new s a emen s o a se ice in he s o ed on ology, Table
6. This se ice ope a es and uns on a de ice wi h ARM926EJ-S pla o m
and he main unc ion is o expo he in o ma ion ha should be added o
he ITS. d s on ology. The se e (expo e ) c ea es a RDF ile ha con ains
14 s a emen s ( iples). This RDF is ma shalled in a s ing and con ains all
he in o ma ion ha will be use ul, in a clien /se e dis ibu ed en i on-
men , so ha he clien can access i . Wi h he new 14 s a emen s added,
he ITS. d s on ology has now 2,896 iple s (added o he o iginal 2,882
25
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
Sal on, G., Buckley, C., 1990. Imp o ing e ie al pe o mance by ele ance
eedback. Jou nal o he Ame ican Socie y o In o ma ion Science 41, 288–
297.
Sileshi, M., Gamback, B., Ma ch 2009. E alua ing clus e ing algo i hms:
Clus e quali y and ea u e selec ion in con en -based image clus e ing.
In: Compu e Science and In o ma ion Enginee ing, 2009 WRI Wo ld
Cong ess on. Vol. 6. pp. 435–441.
Soa es, V., Fa ahmand, F., Rod igues, J., July 2009. A laye ed a chi ec u e
o ehicula delay- ole an ne wo ks. In: Compu e s and Communica-
ions, 2009. ISCC 2009. IEEE Symposium on. pp. 122–127.
Sokal, R., Snea h, P., 1963. P inciples o Nume ical Taxonomy. Books in
biology. W. H. F eeman.
URL h p://books.google.com.py/books?id=3Y4aAAAAMAAJ
Tekli, J., Damiani, E., Chbei , R., Gianini, G., Thi d 2012. Soap p ocessing
pe o mance and enhancemen . Se ices Compu ing, IEEE T ansac ions
on 5 (3), 387–403.
Te ziyan, V., Kayko a, O., Zho ob yukh, D., May 2010. Ubi oad: Seman ic
middlewa e o con ex -awa e sma oad en i onmen s. In: In e ne and
Web Applica ions and Se ices (ICIW), 2010 Fi h In e na ional Con e -
ence on. pp. 295–302.
Thomas, T., an Be kum, E., June 2009. De ec ion o inciden s and e en s
in u ban ne wo ks. In elligen T anspo Sys ems, IET 3 (2), 198–205.
To al, S., To es, M., Ba e o, F., A ahal, M., Sep embe 2010. Cu en
pa adigms in in elligen anspo a ion sys ems. In elligen T anspo Sys-
ems, IET 4 (3), 201–211.
USDOT, 2015.
URL h p://www.i s.do .go /index.h m
Wang, F.-Y., Zeng, D., Yang, L., Oc 2006. Sma ca s on sma oads: An
ieee in elligen anspo a ion sys ems socie y upda e. Pe asi e Compu -
ing, IEEE 5 (4), 68–69.
32

ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
Ye, J., Jana dan, R., Pa k, C. H., Pa k, H., Aug 2004. An op imiza ion
c i e ion o gene alized disc iminan analysis on unde sampled p oblems.
Pa e n Analysis and Machine In elligence, IEEE T ansac ions on 26 (8),
982–994.
Zhai, J., Cao, Y., Chen, Y., Oc 2008. Seman ic in o ma ion e ie al based
on uzzy on ology o in elligen anspo a ion sys ems. In: Sys ems, Man
and Cybe ne ics, 2008. SMC 2008. IEEE In e na ional Con e ence on. pp.
2321–2326.
Zhang, H., Baba , M. A., Tell, P., 2011. Iden i ying ele an s udies in so -
wa e enginee ing. In o ma ion and So wa e Technology 53 (6), 625 – 637,
special Sec ion: Bes pape s om he {APSEC}.
De lis G ego was bo n in Asuncion, Pa aguay, in 1980. He
ecei ed he Bachelo Deg ee in Sys ems Analysis and he Compu e Engi-
nee ing om he Ame ican Uni e si y, Asuncion, Pa aguay, in 2007. Re-
cei ed he M.Sc. and Ph.D. Deg ees in Elec onic, Signal P ocessing and
Communica ions om he Uni e si y o Se ille, Spain, in 2009 and 2013,
espec i ely. He is cu en ly Head o he Labo a o y o Dis ibu ed Sys-
ems, Enginee ing Facul y o he Na ional Uni e si y o Asuncion (FIUNA),
Pa aguay. His esea ch in e es ocuses on he applica ion o In elligen
T anspo a ion Sys ems (ITS). In e ope abili y in Senso Ne wo ks, Embed-
ded Sys ems and Ins umen a ion Sys ems.
Se gio To al ecei ed he M.Sc. and Ph.D. deg ees in elec-
ical and elec onic enginee ing om he Uni e si y o Se ille, Se ille, Spain,
in 1995 and 1999, espec i ely. He is cu en ly a Full P o esso wi h he De-
33
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
pa men o Elec onic Enginee ing, Uni e si y o Se ille. His ecen esea ch
in e es s include senso ne wo ks and in elligen anspo sys ems. P o .
To al was a ecipien o he Bes Pape Awa ds om he IEEE TRANS-
ACTIONS ON INDUSTRIAL ELECTRONICS in 2009 and Ins i u ion o
Enginee ing and Technology Elec ic Powe Applica ions in 2010-2011.
Te esa A iza was bo n in Cadiz, Spain, in 1968. She e-
cei ed he M.S. and Ph.D. deg ees in Compu ing Science om he Uni e -
si y o Se ille, Spain, in 1991 and 2000, espec i ely. She is cu en ly a
ull P o esso wi h he Depa men o Telema ic Enginee ing, US. He main
esea ch in e es s include eal- ime and dis ibu ed sys ems, middlewa es,
in elligen anspo a ion sys ems, embedded ope a ing sys ems and heal h
applica ions.
Fede ico Ba e o ecei ed he M.Sc. and Ph.D. deg ees in
elec ical and elec onic enginee ing om he Uni e si y o Se ille, Se ille,
Spain, in 1992 and 1998, espec i ely.
In 1992, he joined he Elec onic Enginee ing Depa men , Uni e si y o
Se ille, whe e he is cu en ly an Associa e P o esso . His ecen in e es s
include senso ne wo ks and con ol o mul iphase ac d i es.
D . Ba e o was a ecipien o Bes Pape Awa ds om he IEEE TRANS-
ACTIONS ON INDUSTRIAL ELECTRONICS in 2009 and IET Elec ic
Powe Applica ions in 2010-2011.
34
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
Ra´ul G ego was bo n in Asuncion, Pa aguay, in 1979. He
ecei ed he M.Sc. and Ph.D. deg ees om he Uni e si y o Se ille, Spain, in
2006 and 2010 espec i ely. He joined Facul y o Enginee ing o he Na ional
Uni e si y o Asuncion, Pa aguay, in Feb ua y 2009. Since 2012, he is he
Head o he Labo a o y o Powe and Con ol Sys ems, Enginee ing Facul y
o he Na ional Uni e si y o Asuncion (FIUNA). P o . G ego ecei ed he
Bes Pape Awa d om he IEEE TRANSACTIONS ON INDUSTRIAL
ELECTRONICS in 2009, and he Bes Pape Awa d om he Ins i u ion
o Enginee ing and Technology ELECTRIC POWER APPLICATIONS, in
2012.
Jo ge Rodas was bo n in Asuncion (Pa aguay) in 1984.
He ecei ed he Elec onics Enginee Deg ee om he Enginee ing Facul y
o Na ional Uni e si y o Asuncion in 2009. He ecei ed his Mas e ’s de-
g ee in 2012 in Signal P ocessing Applica ions o Communica ions om he
Uni e si y o Vigo (Spain). Since sep embe 2011 he is wi h he Labo a o y
o Powe and Con ol Sys em, Enginee ing Facul y o Na ional Uni e si y
o Asuncion. In 2013 he ob ained a Mas e ’s deg ee in Elec onics, Signal
P ocessing and Communica ions om he Uni e si y o Se ille (Spain). He
is a ecipien o he Fundaci´on Ca olina Pos g adua e Schola ship Awa d o
his PhD s udy.
Ma io A zamendia ecei ed his bachelo deg ee in Elec i-
cal Enginee ing om he Uni e si y o B asilia (B azil) in 2002 and his mas e
35
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
deg ee in Elec onic Enginee ing om Mie Uni e si y (Japan) in 2009. F om
2009 un il 2013 he wo ked as a p ojec leade a he Au oma ion and Con-
ol Inno a ion Cen e (CIAC) o he I aipu Technological Pa k. In 2013
he joined he Facul y o Enginee ing o he Na ional Uni e si y o Asun-
cion as a esea che and since 2014 he is coo dina o o he Labo a o y o
Dis ibu ed Sys ems. His esea ch in e es s include embedded sys ems and
wi eless senso ne wo ks.
36
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
Highligh s
1. We p opose a me hodology o build on ology’s in he domain o ITS (In elligen T anspo a ion Sys ems)
conside ing IR (In o ma ion Reco e y) and SLR (Sys ema ic Li e a u e Re iew).
2. Two new me hods ha e been p oposed: DPK (Disc imina ion o Pa ag aphs wi h Keywo ds) and IRWDP
(Re ie al wi h Weigh ed Da a in Pa ag aphs).
3. The me hods p oposed allow educing he sample size o he s udy.
4. Much in o ma ion i ele an has been disca ded, achie ing g ea e pe o mance in he on ology
cons uc ion.
5. The me hodologies p oposed can be used o build on ologies in any domains.