P ac ice Pape
Recommended ci a ion: Yu, T.-J., Renaud-Assema , I., Beie , S., & Li, D. D.
(2025). Au oma ed Plagia ism De ec ion in CAD Modelling Cou ses:
Ca ego isa ion o Academic Misconduc . In Kangaslampi, R., Langie, G., Jä inen,
H.-M., & Nagy, B. (Eds.), SEFI 53 d Annual Con e ence. Eu opean Socie y o
Enginee ing Educa ion (SEFI), Tampe e, Finland. DOI:
10.5281/zenodo.17631406.
This Con e ence Pape is b ough o you o open access by he 53 d Annual Con e ence
o he Eu opean Socie y o Enginee ing Educa ion (SEFI) a Tampe e Uni e si y in
Tampe e, Finland. This wo k is licensed unde a C ea i e Commons
A ibu ion-NonComme cial-Sha e Alike 4.0 In e na ional License.
AUTOMATED PLAGIARISM DETECTION IN CAD-MODELLING
COURSES: CATEGORISATION OF ACADEMIC MISCONDUCT
T.-J. Yu a, I. Renaud-Assema b, S. Beie c, D. D. Li d,
1
a School o Mechanical and Manu ac u ing Enginee ing,
The Uni e si y o New Sou h Wales (UNSW Sydney), Sydney, Aus alia,
h ps://o cid.o g/0009-0003-1293-218X
b School o Mechanical and Manu ac u ing Enginee ing,
The Uni e si y o New Sou h Wales (UNSW Sydney), Sydney, Aus alia,
h ps://o cid.o g/0000-0001-9960-7950
c School o Mechanical and Manu ac u ing Enginee ing,
The Uni e si y o New Sou h Wales (UNSW Sydney), Sydney, Aus alia,
h ps://o cid.o g/0000-0002-9815-108X
d School o Mechanical and Manu ac u ing Enginee ing,
The Uni e si y o New Sou h Wales (UNSW Sydney), Sydney, Aus alia,
h ps://o cid.o g/0000-0002-6630-5515
Con e ence Key A eas: Digi al ools and AI in enginee ing educa ion, Quali y
assu ance and acc edi a ion o enginee ing educa ional p og ams
Keywo ds: academic in eg i y, CAD, plagia ism de ec ion, SOLIDWORKS
ABSTRACT
P o iciency in c ea ing h ee-dimensional (3D) models using compu e -aided design
(CAD) so wa e is essen ial o enginee ing s uden s. Howe e , plagia ism in CAD-
modelling assessmen s is a g owing conce n in enginee ing educa ion, as mos
de ec ion ools a e de eloped o ex -based submissions. This pape in oduces an
au oma ed plagia ism de ec ion ool designed o CAD assignmen s on he
SOLIDWORKS pla o m (Dassaul Sys èmes). Resul s demons a e ha he ool
enhances de ec ion pe o mance compa ed o exis ing solu ions, imp o ing e iciency
and accu acy in iden i ying po en ial cases o plagia ism. While i s eamlines he
de ec ion p ocess, human judgmen emains c ucial o esol ing ambiguous cases.
Addi ionally, alse posi i es can slow he e iew p ocess, and ecommenda ions a e
made o e ining algo i hms o imp o e eliabili y and accu acy.
1
Co esponding Au ho
D. D. Li
da [email protected]
1 INTRODUCTION
Upholding academic in eg i y is a undamen al p io i y in highe educa ion, ye
main aining i has become inc easingly challenging wi h he ansi ion o online
assessmen (Beie e al., 2022). Pa e ns in plagia ism a e unp edic able (Le ine &
Pazde nik, 2018), especially in he wake o he COVID-19 pandemic (Eshe , 2024);
a ying ac oss assessmen ypes and lea ning en i onmen s. Compu e -aided design
(CAD) assessmen s u he ampli y hese challenges due o he lack o dedica ed
plagia ism de ec ion ools, making e ec i e moni o ing and p e en ion mo e di icul .
So wa e-based plagia ism de ec ion me hodologies signi ican ly imp o e bo h
accu acy and e iciency compa ed o manual p ocesses (Johnson, 2018; Ki s ukas,
2018; Mo is, 2019). While plagia ism de ec ion ools ha e long been a ailable,
adap ing hem o non- ex -based echnical and enginee ing iles p esen s unique
challenges. Success ul implemen a ions exis in p og amming assignmen s, wi h
ools such as JPlag (P echel & Malpohl, 2003) and MOSS (Measu emen O
So wa e Simila i y) (Schleime e al., 2003). Despi e hese ad ances, he e emains
a signi ican gap in he a ailabili y o ools o de ec ing plagia ism in 3D CAD solid
models. P o iciency in CAD is an essen ial skill in enginee ing and indus ial design,
as i enables he c ea ion o p ecise 3D models o design and manu ac u e.
Recen esea ch highligh s he limi ed a ailabili y o plagia ism de ec ion solu ions o
3D CAD models, pa icula ly o SOLIDWORKS (Dassaul Sys èmes), whe e
specialised ools emain sca ce (Ki s ukas, 2018; Mo is, 2019). Many exis ing
app oaches exhibi no able limi a ions. Fo ins ance, while Johnson’s MMEM
p og am success ully ga he ed ea u e me ada a om pa and assembly iles, he
me ada a compa ison was pe o med manually. This eliance on manual compa ison
in ol ed a esou ce-in ensi e e iew o ou pu s, which could lead o po en ially alse
nega i e esul s (Johnson, 2018). Ga land Indus ies’ G ade wo ks, a CAD-g ading
p og am, also p o ides plagia ism de ec ion unc ionali ies. Howe e , i elies solely
on ile-le el me ada a which only includes he use name o he s uden who las
sa ed he ile, a da a eco d ha changes whene e a new use sa es he ile,
limi ing i s e ec i eness in gauging o iginali y (Ga land, 2023). Fu he mo e, Wang e
al. employed ea u e-le el me ada a analysis, bu hei app oach elied on a limi ed
se o plagia ism lags, es ic ing de ec ion sensi i i y and making alse posi i e
il e ing mo e challenging (Wang e al., 2017). Addi ionally, unlike Siemens NX
(Mo is, 2019), SOLIDWORKS iles do no possess a unique ile ID, which
necessi a es a mo e comp ehensi e examina ion o ea u e s uc u es o p ope ies
be ween submissions, a he han me ely compa ing a singula pa ID.
This esea ch aims o c ea e an au oma ed ool o de ec ing plagia ism speci ically
ailo ed o CAD-based assignmen s in SOLIDWORKS, a p emie solid modelling
CAD so wa e ex ensi ely used in bo h indus ial and academic se ings. The ool will
be designed p ima ily o de ec ins ances o plagia ism h ough he di ec exchange
o pa ially o wholly comple e iles, he mos common and conce ning me hod o
plagia ism. This ool aims o signi ican ly enhance he e iciency o he assessmen
ma king p ocess, imp o e he accu acy o plagia ism de ec ion, and ein o ce
academic in eg i y wi hin enginee ing educa ion by p o iding in eg a ed unc ionali ies
o ma ke s, including p ocessing assembly iles, au oma ed compa isons o ile
me ada a o e alua e he p obabili y o plagia ism, and conduc ing ea u e-le el
me ada a compa isons. To maximise accessibili y, he ool will be made open sou ce
once planned enhancemen s, including g ea e cus omisabili y o educa o s, a e
inalised.
This pape i s examines he me ada a s uc u e o SOLIDWORKS iles, which o ms
he ounda ion o ou plagia ism de ec ion app oach. The ool iden i ies simila i ies
indica i e o po en ial misconduc by compa ing me ada a ac oss s uden
submissions. Nex , he design, logic, and ope a ional p inciples o he ool a e
de ailed, explaining how i au oma es and s eamlines he de ec ion p ocess. Finally,
ool’s pe o mance is e alua ed using assignmen submissions om a ecen s uden
coho , demons a ing i s e ec i eness in iden i ying plagia ism e icien ly and
accu a ely.
2 METHODOLOGY
2.1 Model and ea u e p ope ies in SOLIDWORKS
The e a e wo common 3D CAD ‘Model’ ypes ha can be c ea ed in SOLIDWORKS:
a ‘Pa ’ model and an ‘Assembly’ model. These models possess h ee ead-only
e sion his o y a ibu es: “C ea ed” (da e), “Las Sa ed” (da e), and “Las Sa ed by”
(use name). The “C ea ed” (da e) a ibu e is ead-only and canno be al e ed by he
use . The “Las Sa ed” (da e), and “Las Sa ed by” (use name) upda es upon each
ile-sa e ope a ion.
A ‘Pa ’ model c ea ed in SOLIDWORKS comp ises a collec ion o de ining geome ic
‘Fea u es’, such as boss-ex usions, e ol e-ex usions, cu s, holes, ille s, and
pa e ns, alongside he wo-dimensional ske ches u ilised o gene a e hese 3D
geome ies. The ‘Fea u es’ a e lis ed in he Fea u e Manage sideba o
SOLIDWORKS in he sequence o hei c ea ion, as illus a ed in Figu e 1.
Fig. 1. The ea u e manage sideba and ea u e p ope ies dialogue in SOLIDWORKS.
Simila o ‘Models’, each o hese ‘Fea u es’ also con ains a ibu es including i s
name, desc ip ion, and he h ee mos impo an a ibu es o plagia ism de ec ion,
namely, “C ea ed By” (use name), “Da e C ea ed”, and “Da e Las Modi ied”. I is
essen ial o emphasise ha he “C ea ed By” and “Da e C ea ed”, ields a e also
ead-only and canno be al e ed, simila o “C ea ed” (da e) a he ‘Model’ le el.
In addi ion o he geome ic ‘Fea u es’ in a ‘Pa ’ model, an ‘Assembly’ model
con ains a special ype o ‘Fea u e’ called ‘Ma es’, which a e use -de ined cons ain s
be ween wo pa s. Since ma es a e a subse o he ea u e ype, hey sha e he
same h ee e sion his o y a ibu es as hose ound in geome ic ea u es.
2.2 P og am o e iew
The ool was de eloped as a s andalone applica ion independen o SOLIDWORKS,
capable o in e ac ing wi h SOLIDWORKS, Mic oso Excel, and ile sys ems h ough
hei co esponding applica ion p og amming in e aces (API) and lib a ies. The ool
was w i en in he C# language as a .NET Windows Fo ms applica ion, comple e wi h
a g aphical use in e ace (GUI), as shown in Figu e 2.
Fig. 2. P og am GUI a e iles ha e been success ully p ocessed.
Using his GUI, he ma ke o he assessmen submissions can in ui i ely egis e
s uden s’ names and s uden IDs in o he p og am be o e ile scanning. Each s uden
submission, ei he a ‘Pa ’ o an ‘Assembly’, depending on he assessmen , is
scanned sequen ially. All au ho use names ound in ea u es, pa s, assemblies o
ma es, and a selec ion o he ea lies c ea ion and modi ica ion da es, a e s o ed in a
comp ehensi e da a s uc u e in memo y. A e all he iles a e scanned, he p og am
compa es each s uden ’s associa ed ile me ada a wi h o he s uden s’ me ada a,
om he p esen and pas coho s. The da a o pas coho s is accessed ia a mas e
olde o JSON iles, co esponding o each da ase o p e iously p ocessed
assessmen submissions.
A se o 8 plagia ism indica o lags is gene a ed o each s uden , based on speci ic
simila i ies be ween hei submissions and hose o o he s uden s. The i s wo lags
a e “da e ma ch” and “au ho ma ch”, signi ying ha imes amps and/o use names
we e ound o ma ch hose in ano he s uden ’s submission. The nex lag is “da e
be o e assignmen elease”, which is ollowed by he “many au ho s” lag, meaning
ha mul iple use names we e ound wi hin he ile da a. Nex , he e a e h ee lags o
classi y he na u e o he use names p esen in he submission: “use name simila o
s uden ’s eal name”, “common use names”, and “ o eign use name”. The inal lag,
“Fo eign s uden ID”, is ue when a s uden ID- ype use name ha does no belong
o he submi ing s uden is ound. This can be used as an indica o o highly likely
plagia ism.
To use hese plagia ism indica o lags o e ec i e plagia ism de ec ion, eigh
dis inc de ec ion c i e ia we e o mula ed, each ep esen ed by a dis inc , 8-bi bina y
sequence, wi h each bi ep esen ing one plagia ism indica o lag men ioned abo e.
By pe o ming a bi wise compa ison be ween a s uden ’s 8 plagia ism indica o lags
and each o he 8 plagia ism de ec ion c i e ia, e e y s uden can be assigned a
plagia ism isk le el, anging om 0 (indica ing ‘minimal isk’ o plagia ism) o 3
(indica i e o a ‘high isk’ o plagia ism). Please e e o ou p e ious wo k o u he
de ails o his p ocess (Li e al., 2024). This c i e ia-based app oach, which elies on
bi wise compa isons, allows he use o lexibly calib a e he plagia ism de ec ion
logic wi h e e y new da ase o esul s gene a ed by he p og am. Plagia ism
indica o lags and c i e ia can be added o emo ed, and he algo i hms used o lag
gene a ion, such as s uden - o-compu e name ma ching, can be imp o ed o uned
acco ding o he design o he assessmen . This adap abili y enables he use o
educe alse posi i es u he and enhance he iden i ica ion o plagia ism cases.
2.3 Analysis o esul s
The p og am au oma ically ou pu s Excel wo kbooks con aining he essen ial
in o ma ion o e e y submission, including he use names, da es, 8 indica o lags,
and he plagia ism isk le el, allowing human ma ke s o e iew la ge ba ches o
submissions e icien ly. Fo example, wi h 300 submissions, his ini ial sc eening
p ocess akes app oxima ely 15 minu es o iden i y alse posi i es, clea cases o
plagia ism, and submissions equi ing u he examina ion. An example o such an
Excel ou pu is shown in Figu e 3. I equi ed, he ma ke may u he examine
submissions by e iewing each submi ed ile’s ull me ada a s o ed by he ool in
local ex iles o by opening he iles in SOLIDWORKS.
In his s udy, se en modes o ca ego ies o academic misconduc a e in oduced o
o e an insigh in o he pa e ns o dishones beha iou obse ed among de ec ed
s uden s, which a e:
• 1 – Sel -plagia ism;
• 2 – P o iding iles o assis ance o pee s;
• 3a – Submi ing wo k ha has been p epa ed ei he in pa , o in ull, wi h o he
s uden s in he same coho ;
• 3b – Submi ing wo k ha has been p epa ed ei he in pa , o in ull, wi h
s uden s no in he cou se, such as pas s uden s o he cou se;
• 3c – Submi ing wo k ha has been downloaded om an online CAD ile lib a y;
• 4 – Con ac chea ing om an ex e nal p o ide ;
• 5 – Rec ea ing ano he s uden 's pa wi hou copying.
Fig. 3. An example o a ypical summa y epo gene a ed by he p og am.
Th ough a ho ough examina ion o submi ed iles and me ada a, he beha iou o
each s uden in ol ed in academic misconduc is placed in o one o hese ca ego ies.
3 RESULTS
A o al o 296 submissions we e ecei ed om 299 s uden s en olled in he 2024
Te m 3 coho o he MMAN1130 Design and Manu ac u ing cou se a UNSW. The
assignmen equi ed each s uden o submi one assembly ile con aining six cus om
pa s and a ew s anda d bea ings and as ene s, esul ing in a ypical o al o 6 o 10
iles pe s uden , which we e s o ed in zipped olde s.
The submissions downloaded om ou Lea ning Managemen Sys em (LMS) we e
di ec ly p ocessed by he p og am au oma ically, which unzipped iles om all
submissions and analysed a o al o 1612 ‘Pa s’, wi hin 301 ‘Assemblies’, in 20
Fi s
Name
Su name ID Au ho Names Ea lies Da e
Da e
Ma ch
Au ho
Ma ch
Old
Da e
Many
Au ho s
Use names like
s uden name
Common
use names
Fo eign
Use name
Fo eign
S uden IDs
Collude
IDs
Risk
S a us
Thel Vadam 2552 A bi e 12.10.2024 13:27 TRUE TRUE
John Halo 0117 john117 17.10.2024 14:06 TRUE TRUE TRUE 1234[A][D] 3
Jacob K. C709 jacobk 10.10.2024 23:07 TRUE
A e y J. 1234 sa ge, john117 17.10.2024 14:06 TRUE TRUE TRUE TRUE 0117[A][D] 3
Jane Ci izen 2009 janec 09.10.2024 09:00 TRUE
minu es and 48 seconds, including he 16 seconds needed o SOLIDWORKS o
launch in he backg ound. The ex a assemblies a e explained by s uden s
mis akenly submi ing mul iple assembly iles. The subsequen c ea ion o a JSON
ile o aw da a s o age and summa y epo s in Excel o ma ook he p og am
a ound 4 seconds, wi h he JSON ile encompassing all submission da a, o alling
667 kiloby es (kB) in size. A summa y o he numbe o plagia ism cases lagged by
he p og am is de ailed in Table 1.
Table 1. Summa y o e dic s on de ec ed cases o plagia ism.
Plagia ism isk le el ( o al numbe
o de ec ed cases by p og am)
Numbe o s uden s
Con i med
plagia ism
Likely alse
posi i e
Clea alse
posi i e
Le el 3: “high isk” (18 cases)
11
1
6
Le el 2: “mode a e isk” (22 cases)
6
3
13
Le el 1: “low isk” (2 cases)
2
0
0
Le el 0: “minimal isk” (2 s uden s)
2
-
-
In o al, 42 cases we e lagged by he p og am wi h a plagia ism p obabili y le el o 1
o highe , o which 19 we e con i med as clea alse posi i es, p ima ily due o issues
a ising om he ile au ho name compa ison p ocesses, such as he inabili y o
ecognise a use name as being simila o he s uden ’s ac ual name (e.g. Thomas s.
om_pc) o cases in ol ing s uden s wi h non-common compu e use names
un ela ed o hei eal name (e.g. John s. mas e chie ). Fou cases wa an ed mo e
de ailed manual examina ion (las ing app oxima ely 20-30 minu es), which
de e mined hem o be alse posi i es as well, and as such hey ha e been classi ied
as "likely alse posi i es".
In addi ion o he emaining 19 cases, wo mo e submissions we e con i med as
plagia ism. Al hough designa ed as 'minimal isk' (le el 0) based on hei own ile
cha ac e is ics, he p og am lagged hem o me ada a ma ches wi h highe - isk
submissions, leading o hei misconduc being con i med, making o a o al o 21
con i med cases o plagia ism. These cases we e hen ca ego ised by he ea lie
discussed “mode o academic misconduc ” in ol ed. O he 21 con i med cases,
eigh we e classi ied as Mode 2, in ol ing p o iding iles o o he assis ance o ellow
s uden s. Nine we e Mode 3a, whe e s uden s submi ed wo k ha included
con ibu ions om o he s. Two cases we e Mode 3b, in ol ing using iles o ea u es
c ea ed by s uden s om p e ious coho s. One case was Mode 4, which in ol ed
con ac chea ing, and one inal case was Mode 5, whe e a s uden ec ea ed
ano he s uden 's wo k wi hou di ec ly copying i . The esul s a e summa ised in
Table 2, along wi h de ails o how he p og am o ma ke iden i ied he cases.
4 DISCUSSION
This inno a i e plagia ism de ec ion ool o SOLIDWORKS CAD models p o ides an
e icien app oach o iden i ying academic misconduc . The ool enabled a single
ma ke o comple e he de ec ion p ocess o 296 submissions wi hin one hou – a
signi ican imp o emen o e p e ious manual me hods, which equi ed a eam o
ou ma ke s wo king o eigh hou s o comple e he same ask. The p e ious
manual app oach was limi ed o inspec ing he ‘Las Sa ed By’ use name a he
‘Model’ le el, making i ime-consuming and ine ec i e a de ec ing sub le o ms o
misconduc . In con as , his au oma ed ool iden i ied plagia ism h ough addi ional
indica o s, such as ma ching imes amps and use names a he ‘Fea u e’ le el, e en
when s uden s used gene ic compu e names – pa e ns ha could easily be
o e looked in a manual e iew. As such, since he p og am eplica es all he key
s eps o a manual e iew, while pe o ming hem wi h signi ican ly g ea e
ho oughness and accu acy, i is unlikely o miss ins ances o plagia ism ha a
manual e iew would de ec , excep in cases in ol ing human e o du ing he inal
p ocessing s age.
Table 2. De ails o all 21 con i med cases o academic misconduc o his assignmen ,
so ed by hei p og am-assigned plagia ism isk le els.
S uden
No.
Risk
le el
Mode o
Misconduc
Desc ip ion
1
3
3a
Use name and imes amp ma ch: Submi ed iles sha ed
by S uden 18.
2
3
3a
Use name and imes amp ma ch: Submi ed iles sha ed
by S uden 19
3
3
2
Sha ed all pa iles, ully assembled, wi h S uden 4.
4
3
3a
Use name and imes amp ma ch: Submi ed iles sha ed
by S uden 3 and eplaced some sha ed pa iles wi h
hei own.
5
3
3a
Use name and imes amp ma ch: sha ed iles wi h and
submi ed iles made by S uden 6.
6
3
3a
Use name and imes amp ma ch: sha ed iles wi h and
submi ed iles made by S uden 5.
7
3
3a
Use name and imes amp ma ch: Submi ed iles sha ed
by S uden 8, wi h added assembly ma es.
8
3
2
Sha ed all pa iles, ully assembled, wi h S uden 7.
9
3
3a
Use name and imes amp ma ch: Submi ed iles sha ed
by S uden 10, wi h added assembly ma es.
10
3
2
Sha ed all pa iles, ully assembled, wi h S uden 9.
11
3
3a
One o he submi ed pa s was c ea ed by S uden 20.
12
2
3b
Use name ma ch: Assembly ma es done by pas s uden .
13
2
4
All ea u es and assembly ma es done by one au ho ,
whose use name ma ches a CAD eelance wi h a
LinkedIn p esence.
14
2
2
Use name ma ch: c ea ed he same pa wice and sen
one o S uden 15, o a oid de ec ion. C ea ed assembly
ma es o S uden 15.
15
2
3a
See S uden 14.
16
2
2
Suspec ed o sending iles o S uden 21 o ac as a
e e ence ile.
17
2
3b
Use name ma ch: All pa s and ea u es made by pas
s uden , while assembly ma es a e o iginal.
18
1
2
Sha ed all pa iles, ully assembled, wi h S uden 1.
19
1
2
Sha ed all pa iles, ully assembled, wi h S uden 2.
20
0
2
Use name ma ch: c ea ed he same pa wice and sen
one o S uden 11, o a oid de ec ion.
21
0
5
Single Use name ma ch: S uden 16's use name was
ound in he assembly p ope ies. This led o a close
examina ion o design decisions, which de e mined ha
al hough no iles we e copied, he design decisions wi hin
he wo s uden s' submissions we e oo simila o be
independen wo k.
The esul s o his s udy highligh se e al signi ican ad ancemen s o e exis ing
app oaches, e ec i ely add essing limi a ions in p io li e a u e. Fo example, unlike
Johnson’s MMEM, which elied on manual me ada a compa ison (Johnson, 2018),
ou ool ully au oma es his p ocess, elimina ing labou -in ensi e e iew s eps and
enhancing scalabili y ac oss la ge s uden coho s. Fu he mo e, Ga land’s
G ade wo ks ool, hough capable o de ec ing plagia ism, elied only on ile-le el
me ada a (Ga land, 2023), which could be easily al e ed when a di e en use sa es
he ile. In con as , ou me hod e ec i ely add esses his limi a ion by scanning and
s o ing all c i ical me ada a o he ‘Fea u e’ le el. Finally, while he me hodology in
Wang e al. u ilises ea u e-le el me ada a collec ion, hei analysis in ol ed ewe
iden i ied plagia ism lags compa ed o ou me hod, which complica es he il e ing o
alse posi i es (Wang e al., 2017). Ou app oach add esses his issue by elying on
mo e plagia ism lags o imp o e de ec ion a es and educe alse posi i es.
Human e alua ion o he p og am’s ou pu s emains essen ial, as 45% o he
iden i ied submissions we e clea alse posi i es. This aligns wi h Ki s ukas’ indings,
whe e 64% o de ec ed cases we e alse posi i es (Ki s ukas, 2018). Mos
misclassi ica ions s emmed om challenges in ca ego ising ile au ho use names.
Since a ma ke ’s in es iga ion is s ill equi ed o con i m cases o plagia ism,
s uden s canno be w ongly inc imina ed wi hou human judgmen . The e is an
inhe en ade-o be ween minimising alse posi i es and main aining su icien
sensi i i y o de ec ac ual misconduc . The sys em’s con igu a ion should e lec he
use ’s ole ance o isk, balancing accu acy and co e age. Fu u e p og am e sions
could inco po a e ad anced name-ma ching algo i hms, such as uzzy ma ching, o
imp o e he accu acy o associa ing s uden names and use names.
The cu en p og am is designed o de ec ins ances o plagia ism p ima ily h ough
he di ec exchange o pa ially o wholly comple e iles. I does no accoun o cases
whe e s uden s independen ly ep oduce ano he s uden ’s wo k wi hou di ec
copying o ile exchange. A solu ion could in ol e analysing he sequence and ypes
o ea u es u ilised in c ea ing CAD models and he me hodologies p oposed he ein
o enhance de ec ion capabili ies. Ne e heless, i is essen ial o no e ha s uden s
imi a ing one ano he ’s CAD wo k is a lea ning expe ience, and his mode o
plagia ism is less signi ican han o he o ms.
5 CONCLUSION
A comp ehensi e plagia ism de ec ion ool o CAD-based assignmen submissions
has been e ec i ely de eloped and implemen ed o iden i y po en ial academic
misconduc o u he in es iga ion p omp ly. The p og am can e icien ly p ocess a
signi ican olume o s uden submissions wi h enhanced accu acy, ho oughness,
and esou ce managemen , signi ican ly educing he demands on educa o s' ime
and esou ces. The ool p o ides a s uc u ed app oach o ini ial plagia ism
iden i ica ion by u ilising me ada a analysis. Howe e , human assessmen emains
essen ial o il e ou alse posi i es and o assess he na u e and se e i y o any
de ec ed plagia ism, ollowing he Uni e si y’s plagia ism policies. The de elopmen
and implemen a ion o he ool aim o suppo educa o s in upholding academic
in eg i y and sa egua ding he epu a ion o educa ional ins i u ions amids cons an ly
changing s uden a i udes owa ds plagia ism; speci ically, collusion wi hin indi idual
assessmen s. The p og am can be e ined based on use eedback o enhance i s
e ec i eness u he .
