Co esponding au ho : Anand Raghu anshi.
Copy igh © 2025 Au ho (s) e ain he copy igh o his a icle. This a icle is published unde he e ms o he C ea i e Commons A ibu ion License 4.0.
AI and p edic i e analy ics in highe educa ion: A sales o ce app oach
Anand Raghu anshi *
Raji Gandhi P oudyogiki Vishwa idyalaya (RGPV), India.
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 3047-3053
Publica ion his o y: Recei ed on 07 Ap il 2025; e ised on 19 May 2025; accep ed on 21 May 2025
A icle DOI: h ps://doi.o g/10.30574/wja .2025.26.2.1958
Abs ac
Highe educa ion ins i u ions inc easingly le e age a i icial in elligence and p edic i e analy ics o enhance s uden
ou comes and ope a ional e iciency. This a icle explo es he implemen a ion o Sales o ce-based p edic i e analy ics
solu ions in academic en i onmen s, ocusing on echnical ounda ions, a chi ec u al componen s, pe sonalized
lea ning pa hways, implemen a ion challenges, and eal-wo ld case s udies. The echnical in as uc u e suppo ing
hese ini ia i es combines sophis ica ed machine lea ning algo i hms wi h di e se da a sou ces o iden i y a - isk
s uden s, pe sonalize lea ning expe iences, and empowe da a-d i en decision-making. Th ough examina ion o
implemen a ions a leading uni e si ies, he a icle demons a es how p ope ly designed p edic i e sys ems deli e
measu able imp o emen s in e en ion, g adua ion a es, and s uden success while p o iding subs an ial e u ns on
in es men . The in eg a ion o ecommenda ion sys ems, adap i e assessmen engines, and lea ning analy ics c ea es
pe sonalized educa ional expe iences, while hough ul implemen a ion s a egies add ess challenges ela ed o da a
in eg a ion, p i acy, model ai ness, and use adop ion.
Keywo ds: A i icial In elligence; Educa ional Technology; P edic i e Modeling; S uden Success; Da a Go e nance
1. In oduc ion
Highe educa ion ins i u ions ace inc easing p essu e o imp o e s uden ou comes, op imize esou ce alloca ion, and
demons a e measu able esul s. In esponse, uni e si ies a e u ning o a i icial in elligence (AI) and p edic i e
analy ics o ans o m hei ope a ional and educa ional app oaches. Sales o ce, wi h i s Educa ion Cloud pla o m, has
eme ged as a leading solu ion p o ide in his space, o e ing powe ul ools ha enable ins i u ions o ha ness da a-
d i en insigh s o s a egic decision-making. This echnical a icle explo es how AI and p edic i e analy ics,
pa icula ly h ough Sales o ce implemen a ions, a e e olu ionizing highe educa ion by iden i ying a - isk s uden s,
pe sonalizing lea ning expe iences, and empowe ing adminis a o s wi h ac ionable in elligence.
The u gency o adop ing ad anced analy ics solu ions is unde sco ed by ecen da a om he esea ch, e ealing ha
he 6-yea g adua ion a e o i s - ime, ull- ime unde g adua es who began seeking a bachelo 's deg ee a 4-yea
deg ee-g an ing ins i u ions in all 2016 was 69 pe cen o e all, wi h p i a e nonp o i ins i u ions achie ing 78 pe cen
compa ed o 62 pe cen a public ins i u ions [1]. These s a is ics highligh he pe sis en challenges ins i u ions ace in
suppo ing s uden s h ough deg ee comple ion, pa icula ly ac oss di e en ins i u ional ypes and demog aphic
g oups.
P edic i e analy ics o e s a solu ion by enabling ea ly iden i ica ion o s uden isk ac o s. As esea ch by Rajni Jindal
and Malaya Du a Bo ah indica es, educa ional analy ics can be e ec i ely employed o p edic s uden g ades wi h 70-
80% accu acy and iden i y s uden s a isk o d opping ou wi h simila p ecision [2]. Thei s udy u he demons a es
ha p edic i e models can signi ican ly enhance s uden success ini ia i es by analyzing his o ical da a pa e ns ac oss
academic pe o mance, engagemen me ics, and demog aphic ac o s o c ea e a ge ed in e en ion s a egies. When
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 3047-3053
3048
implemen ed h ough pla o ms like Sales o ce Educa ion Cloud, hese analy ics-d i en app oaches allow ins i u ions
o mo e beyond eac i e measu es o p oac i e s uden suppo sys ems ha add ess indi idual needs be o e s uden s
each c i ical isk h esholds.
2. The Technical Founda ion o Educa ional P edic i e Analy ics
P edic i e analy ics in highe educa ion elies on sophis ica ed machine lea ning algo i hms applied o di e se da ase s.
These sys ems in eg a e and analyze da a om mul iple sou ces including S uden In o ma ion Sys ems (SIS), Lea ning
Managemen Sys ems (LMS), cou se egis a ion and a endance eco ds, academic pe o mance me ics, engagemen
indica o s (lib a y usage, online ac i i y, e c.), and his o ical s uden ou come da a.
Recen ad ancemen s in educa ional da a mining ha e demons a ed he c i ical impo ance o comp ehensi e da a
in eg a ion. Acco ding o Joshua Pa ick Ga dne and Ch is ophe B ooks, whose sys ema ic e iew analyzed 91
p edic i e modeling s udies in educa ional con ex s, only 42% o s udies inco po a ed da a om mul iple ins i u ional
sys ems, despi e e idence ha mul i-sou ce models signi ican ly ou pe o m single-sou ce app oaches [3]. Thei
analysis e ealed ha models inco po a ing bo h SIS and LMS da a achie ed classi ica ion imp o emen s anging om
5-15% compa ed o models using only one da a sou ce. Fu he mo e, hey iden i ied ha empo al ea u es
( ep esen ing how s uden beha io s change o e ime) we e among he mos p edic i e a iables ye we e
unde u ilized in only 23% o he s udies examined.
Sales o ce's Eins ein AI laye p ocesses hese di e se da ase s using se e al key echniques: supe ised lea ning models
o classi ica ion p oblems, eg ession algo i hms o o ecas ing con inuous a iables, Na u al Language P ocessing o
sen imen analysis, clus e ing echniques o s uden segmen a ion, and neu al ne wo ks o iden i ying complex
ela ionships in educa ional da a.
These echniques build upon g oundb eaking wo k by Jiang e al., who demons a ed ha ensemble models signi ican ly
ou pe o m single algo i hms in educa ional con ex s [4]. Thei s udy examining 32,538 eco ds o s uden cou se
in e ac ions ound ha andom o es models achie ed an AUC o 0.802 when p edic ing s uden s a isk o ailing
cou ses, compa ed o 0.731 o logis ic eg ession and 0.688 o decision ees. Mos no ably, hei esea ch es ablished
ha p edic ion accu acy imp o es d ama ically when models inco po a e bo h s a ic s uden cha ac e is ics and
dynamic beha io al ea u es, wi h weekly clicks eam da a om lea ning managemen sys ems imp o ing p edic i e
pe o mance by 7.7% compa ed o models using only demog aphic and his o ical academic da a.
The Eins ein Analy ics pla o m u ilizes hese algo i hms wi hin a scalable cloud a chi ec u e, allowing ins i u ions o
p ocess massi e da ase s while main aining FERPA compliance h ough obus secu i y p o ocols. This a chi ec u al
app oach e lec s bes p ac ices iden i ied in he li e a u e, which emphasize ha educa ional p edic i e models mus
balance echnical sophis ica ion wi h in e p e abili y o be ac ionable o educa ional s akeholde s while espec ing
s uden p i acy conce ns.
Table 1 P edic i e Model Accu acy in Iden i ying A -Risk S uden s: Algo i hm Compa ison [3, 4]
Algo i hm Type
AUC Sco e
Rela i e Pe o mance (%)
Logis ic Reg ession
0.731
91.1
Decision T ees
0.688
85.8
Single-Sou ce Models (A e age)
0.650
81.0
Mul i-Sou ce Models (A e age)
0.748
93.3
Models wi h S a ic Fea u es Only
0.695
86.7
Models wi h S a ic + Dynamic Fea u es
0.748
93.3
3. Implemen a ion A chi ec u e o Ea ly Wa ning Sys ems
Ea ly wa ning sys ems (EWS) ep esen one o he mos impac ul applica ions o p edic i e analy ics in highe
educa ion. A ypical Sales o ce-based EWS implemen a ion ollows a comp ehensi e echnical a chi ec u e ha
in eg a es mul iple sys ems o enable imely in e en ions.
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 3047-3053
3049
The Da a In eg a ion Laye es ablishes API-d i en connec ions o S uden In o ma ion Sys ems (SIS), Lea ning
Managemen Sys ems (LMS), and o he sys ems, wi h ETL p ocesses handling da a no maliza ion and quali y assu ance.
Acco ding o Kimbe ly E. A nold, Ma hew D. Pis illi's seminal wo k on Cou se Signals a Pu due Uni e si y, his
in eg a ion laye enabled hei sys em o analyze o e 20 di e en da a poin s pe s uden d awn om mul iple
ins i u ional sys ems [5]. Thei implemen a ion demons a ed ha e ec i e da a in eg a ion suppo s ea ly
iden i ica ion o a - isk s uden s, wi h in e en ions possible as ea ly as he second week o cou ses.
The Da a Lake/Wa ehouse p o ides a s uc u ed eposi o y o his o ical and eal- ime da a s o age. The P edic i e
Engine le e ages Eins ein Disco e y models ained on ins i u ional da a o iden i y isk pa e ns using academic
pe o mance indica o s, beha io al me ics, and demog aphic a iables. Kimbe ly E. A nold and Ma hew D. Pis illi's
esea ch showed ha his app oach enabled Cou se Signals o achie e signi ican imp o emen s in s uden ou comes,
wi h cou ses u ilizing he sys em showing a 10% inc ease in A and B g ades and a 6.41% dec ease in D and F g ades
compa ed o cou ses wi hou he sys em [5].
The Visualiza ion Laye employs Ligh ning-based dashboa ds p esen ing isk assessmen s wi h d ill-down capabili ies,
while he In e en ion Managemen Sys em p o ides wo k low au oma ion ou ing ale s o app op ia e pe sonnel.
Resea ch by Baeple and Mu doch e ealed ha e ec i e isualiza ion and in e en ion sys ems inc eased ad iso
capaci y o manage s uden cases by app oxima ely 30%, enabling mo e pe sonalized suppo [6]. Thei s udy o
educa ional echnology implemen a ions demons a ed ha sys ems p o iding bo h isk iden i ica ion and s uc u ed
in e en ion capabili ies achie ed signi ican ly highe adop ion a es among acul y and ad iso s.
The Feedback Loop Mechanism comple es he a chi ec u e by acking in e en ion e ec i eness and model
pe o mance. Baeple and Mu doch's analysis showed ha ins i u ions implemen ing sys ema ic ou come acking and
pe o mance moni o ing saw p og essi e imp o emen in model accu acy, wi h hi d-gene a ion implemen a ions
co ec ly iden i ying 85% o a - isk s uden s compa ed o 71% in ini ial deploymen s [6]. This eedback componen is
pa icula ly c i ical, as hei esea ch indica ed ha in e en ions igge ed by he sys em led o an a e age
imp o emen o 1.8 pe cen age poin s in cou se comple ion a es du ing he i s yea o implemen a ion.
This in eg a ed a chi ec u e enables nea eal- ime iden i ica ion o a - isk s uden s wi h accu acy a es ypically
exceeding 80% when p ope ly implemen ed and calib a ed o ins i u ion-speci ic pa e ns.
4. Pe sonalized Lea ning Pa h Op imiza ion
Beyond isk iden i ica ion, p edic i e analy ics enables indi idualized lea ning pa hways. The echnical implemen a ion
ypically in ol es se e al in eg a ed componen s wo king oge he o c ea e pe sonalized educa ional expe iences.
Recommenda ion Sys ems employ collabo a i e il e ing algo i hms o iden i y op imal cou se sequences based on
simila s uden ou comes. Acco ding o Ami Hossein Nabizadeh e al., hese sys ems can be ca ego ized in o con en -
based il e ing (CBF), collabo a i e il e ing (CF), and hyb id app oaches, wi h collabo a i e il e ing demons a ing
supe io pe o mance in educa ional con ex s when su icien da a is a ailable [7]. Thei comp ehensi e su ey
e ealed ha CF-based sys ems ha e shown imp o emen s in s uden pe o mance anging om 9-15% when p ope ly
implemen ed. They also ound ha ma ix ac o iza ion echniques ou pe o m neighbo hood me hods in 73% o
educa ional ecommenda ion scena ios due o hei abili y o disco e la en ac o s in s uden lea ning pa e ns.
Adap i e Assessmen Engines u ilize i em esponse heo y models ha adjus con en di icul y based on demons a ed
mas e y. Ryan S. Bake e al., esea ch on adap i e lea ning sys ems shows ha hese engines can educe he ime
equi ed o s uden assessmen by 25-30% while simul aneously imp o ing lea ning ou comes by collec ing mo e
p ecise in o ma ion abou s uden knowledge s a es [8]. Thei analysis e ealed ha adap i e sys ems implemen ing
compu e ized adap i e es ing (CAT) demons a ed imp o ed measu emen p ecision wi h up o 50% ewe i ems
compa ed o adi ional ixed- o m assessmen s.
Lea ning Analy ics p o ides empo al analysis o engagemen pa e ns o iden i y op imal s udy echniques o
indi idual s uden s. The esea ch by Ryan S. Bake e al., demons a es ha analysis o s uden LMS in e ac ion da a can
iden i y a - isk s uden s wi h accu acy a es o 70-90% by he ou h week o cou ses, enabling imely in e en ions [8].
Thei wo k shows ha sys ems analyzing empo al pa e ns in lea ning ac i i ies can dis inguish be ween e ec i e and
ine ec i e lea ning beha io s wi h 75-85% accu acy.
In eg a ion Poin s es ablish APIs connec ing Sales o ce wi h adap i e lea ning pla o ms, while Ou come Op imiza ion
algo i hms balance comple ion ime, mas e y le el, and esou ce u iliza ion. Ami Hossein Nabizadeh e al. no e ha
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 3047-3053
3050
se ice-o ien ed a chi ec u e implemen a ions wi h REST APIs ha e become he dominan in eg a ion app oach, wi h
adop ion a es exceeding 80% in ecen educa ional echnology implemen a ions [7].
These sys ems le e age Sales o ce's Jou ney Builde o o ches a e pe sonalized communica ion lows, deli e ing
a ge ed con en and guidance a op imal in e als. Technical in eg a ion is achie ed h ough REST APIs wi h
au hen ica ion handled ia OAu h 2.0 p o ocols, ensu ing secu e da a exchange be ween lea ning pla o ms and he
Sales o ce co e. As highligh ed by Ryan S. Bake e al., his in eg a ion app oach suppo s an a e age o 21-32 dis inc
communica ion ouchpoin s pe s uden pe e m while main aining da a secu i y and compliance wi h educa ional
p i acy equi emen s [8].
Table 2 E ec i eness Compa ison o Educa ional Technology Componen s in Pe sonalized Lea ning [7, 8]
Technology Componen
Pe o mance Me ic
Value
(%)
Compa ison Poin
Compa ison
Value (%)
Collabo a i e Fil e ing
Recommenda ion Sys ems
S uden Pe o mance
Imp o emen
9-15
Neighbo hood
Me hods
27%
Ma ix Fac o iza ion Techniques
Success Ra e in
Educa ional Scena ios
73
Neighbo hood
Me hods
27%
REST API In eg a ion
App oaches
Adop ion Ra e in
Educa ional Technology
80
O he In eg a ion
Me hods
20%
5. Technical Challenges and Solu ions in Implemen a ion
Implemen ing AI-d i en analy ics in educa ional se ings p esen s se e al echnical challenges ha equi e hough ul
solu ions o ensu e success ul adop ion and e hical use.
Da a silos ac oss depa men s ep esen a undamen al obs acle o e ec i e analy ics implemen a ion. Resea ch by
Ca ie Klein e al., examining ou highe educa ion ins i u ions ound ha 65% o acul y epo ed di icul ies in
accessing da a needed o lea ning analy ics due o o ganiza ional silos [9]. Thei s udy e ealed ha success ul
implemen a ions add essed his challenge h ough comp ehensi e in eg a ion s a egies, wi h one ins i u ion epo ing
ha API-based in eg a ion app oaches educed da a access ba ie s by c ea ing s anda dized access poin s ac oss
p e iously isola ed sys ems. Ins i u ions implemen ing canonical da a models demons a ed pa icula success in
c ea ing common unde s anding o da a elemen s ac oss di e se s akeholde s, wi h Ca ie Klein esea ch showing
imp o ed communica ion be ween echnical and non- echnical s a when s anda dized da a de ini ions we e
es ablished.
P i acy compliance wi h egula ions such as FERPA and GDPR p esen s ano he c i ical challenge. Ca ie Klein e al.,
iden i ied p i acy conce ns as one o he mos signi ican ba ie s o analy ics adop ion, wi h 69% o acul y exp essing
ese a ions abou s uden da a use [9]. Thei esea ch ound ha implemen a ions inco po a ing ield-le el
enc yp ion, ole-based access con ols, and da a masking signi ican ly inc eased s akeholde com o le els, wi h one
ins i u ion epo ing ha anspa en p i acy amewo ks inc eased acul y willingness o pa icipa e in analy ics
ini ia i es by 40%.
Model bias and ai ness conce ns mus be add essed o a oid ein o cing exis ing inequi ies. René F. Kizilcec and Hansol
Lee ex ensi e e iew o algo i hmic ai ness in educa ional echnology iden i ied ha p edic ion dispa i ies commonly
exis ac oss demog aphic g oups when using s anda d algo i hms [10]. Thei esea ch demons a ed ha ai ness-
awa e app oaches can signi ican ly educe hese dispa i ies, wi h one s udy showing ha adjus ed models educed
accu acy gaps be ween demog aphic g oups om 13.0% o jus 1.8%. They emphasize ha egula bias audi s a e
essen ial, as model pe o mance dispa i ies can eme ge o e ime e en when ini ial aining da a is balanced.
Algo i hm explainabili y is essen ial o building us wi h educa ional s akeholde s. Resea ch on SHAP alues and
o he explana ion echniques has shown ha acul y engagemen wi h analy ics inc eases subs an ially when models
p o ide in e p e able p edic ions [10]. Scale and pe o mance challenges in ensi y as implemen a ions expand, wi h
Sales o ce's asynch onous p ocessing capabili ies o e ing solu ions o main aining esponsi eness du ing high-
olume ope a ions.
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 3047-3053
3051
Use adop ion emains pe haps he mos pe sis en challenge. Ca ie Klein e al., esea ch e ealed ha success ul
implemen a ions achie ed 70-80% adop ion a es by ocusing on p og essi e in e ace design wi h con ex ual guidance
[9].
Ins i u ions success ully na iga e hese challenges h ough phased implemen a ions, beginning wi h ocused use cases
be o e expanding o en e p ise-wide deploymen s. C i ical o success is es ablishing obus da a go e nance
amewo ks ha balance analy ical capabili y wi h e hical conside a ions, wi h Ca ie Klein e al., demons a ing ha
go e nance s uc u es signi ican ly impac bo h adop ion a es and e hical implemen a ion [9].
Table 3 Technical Challenges in Educa ional Analy ics Implemen a ion: Impac and Mi iga ion Me ics [9, 10]
Challenge Ca ego y
Solu ion App oach
Da a Silos
API-based In eg a ion
Canonical Da a Models
P i acy Conce ns
Field-le el Enc yp ion
Role-based Access Con ols
Da a Masking
T anspa en P i acy F amewo ks
Model Bias
Fai ness-awa e App oaches
Regula Bias Audi s
Algo i hm Explainabili y
SHAP Values
Use Adop ion
P og essi e In e ace Design
Con ex ual Guidance
O e all Implemen a ion
Phased Implemen a ion App oach
Robus Da a Go e nance
6. Case S udies: Quan i iable Impac Me ics
Se e al ins i u ions ha e demons a ed measu able success wi h Sales o ce-based p edic i e analy ics
implemen a ions, p o iding compelling e idence o he impac o hese echnologies on s uden ou comes.
A izona S a e Uni e si y has eme ged as a leade in applying p edic i e analy ics o imp o e s uden success. Acco ding
o esea ch by Kelli Bi d ASU inc eased e en ion a es by 12% h ough ea ly in e en ion s a egies implemen ed ia
hei eAd iso sys em [11]. Thei implemen a ion educed ime- o-deg ee by 0.8 semes e s on a e age while achie ing
83% accu acy in p edic ing a - isk s uden s by mid e m. As Kelli Bi d no e in hei comp ehensi e e iew, ASU's success
s ems om hei sys ema ic app oach o in e en ion, wi h hei p edic i e models analyzing pa e ns ac oss cou se
pe o mance, engagemen me ics, and demog aphic ac o s o igge imely suppo mechanisms be o e s uden s
each c i ical isk h esholds.
The Uni e si y o Ken ucky demons a es ano he success ul implemen a ion, le e aging Eins ein Analy ics o p ocess
o e 700 a iables pe s uden . Kelli Bi d’s esea ch documen s how his comp ehensi e app oach enabled Ken ucky
o inc ease 6-yea g adua ion a es by 8.1 pe cen age poin s while educing achie emen gaps o unde ep esen ed
g oups by 6.2% [11]. Thei analysis highligh s ha Ken ucky's implemen a ion success was la gely due o hei ocus
on wha hey e m "ac ionable in elligence"—ensu ing ha p edic i e insigh s we e di ec ly connec ed o speci ic
in e en ion pa hways managed h ough he Sales o ce pla o m.
Geo gia S a e Uni e si y's implemen a ion ep esen s one o he mos ex ensi ely documen ed success s o ies in
educa ional analy ics. Kimbe ly E. A nold, S e en Lonn and Ma hew D. Pis illi's analysis o GSU's app oach e eals how
hei sys em iden i ies mo e han 800 dis inc isk ac o s o gene a e 52,000 p oac i e in e en ions annually [12].
Thei esea ch documen s how GSU inc eased g adua ion a es by 23% o e a i e-yea pe iod h ough his sys ema ic
app oach. Kimbe ly E. A nold, S e en Lonn and Ma hew D. Pis illi speci ically no e ha GSU's implemen a ion
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 3047-3053
3052
demons a es he impo ance o ins i u ional eadiness o analy ics adop ion, wi h hei Lea ning Analy ics Readiness
Ins umen (LARI) assessmen showing ha GSU sco ed pa icula ly high on leade ship commi men and da a cul u e
dimensions.
These case s udies demons a e he angible impac o well-implemen ed p edic i e sys ems. Kimbe ly E. A nold, S e en
Lonn and Ma hew D. Pis illi's esea ch on analy ics eadiness iden i ies se e al c i ical success ac o s common ac oss
high-pe o ming implemen a ions, including clea go e nance s uc u es, s akeholde engagemen , and echnical
in as uc u e [12]. Thei analysis sugges s ha ins i u ions wi h high eadiness sco es achie e signi ican ly be e
ou comes om analy ics implemen a ions, wi h ROI calcula ions showing e u ns o $3-7 o e e y dolla in es ed in
analy ics in as uc u e when accoun ing o inc eased e en ion and educed adminis a i e o e head.
7. Conclusion
The u u e o AI-d i en educa ional analy ics con inues o e ol e h ough deepe in eg a ion o eme ging echnologies
and me hodologies. Edge compu ing will enable mo e immedia e analy ics di ec ly a s uden in e ac ion poin s, while
ede a ed lea ning app oaches add ess p i acy conce ns by aining models wi hou cen alizing sensi i e in o ma ion.
Na u al language gene a ion will ans o m insigh communica ion h ough au oma ed na a i e c ea ion o a ious
s akeholde s. The ounda ion es ablished by cu en Sales o ce implemen a ions p o ides a amewo k o u u e
expansion, hough ins i u ions mus main ain a balance be ween echnological capabili ies and e hical conside a ions.
P edic i e models unc ion op imally when se ing as ools o human decision-make s a he han au oma ed
eplacemen s. When implemen ed wi h app op ia e go e nance s uc u es, AI-powe ed p edic i e analy ics ep esen s
a ans o ma i e o ce in highe educa ion, enabling ins i u ions o ul ill educa ional missions mo e e ec i ely h ough
da a-in o med app oaches o s uden success, c ea ing mo e equi able and pe sonalized lea ning en i onmen s while
op imizing ins i u ional esou ces.
Re e ences
[1] Ke Wang e al., "Repo on he Condi ion o Educa ion 2024," Na ional Cen e o Educa ion S a is ics a IES, 2024.
[Online]. A ailable: h ps://nces.ed.go /pubs2024/2024144.pd
[2] Rajni Jindal and Malaya Du a Bo ah, "P edic i e Analy ics in a Highe Educa ion Con ex ," Resea chGa e, 2015.
[Online]. A ailable:
h ps://www. esea chga e.ne /publica ion/280931592_P edic i e_Analy ics_in_a_Highe _Educa ion_Con ex .
[3] Joshua Pa ick Ga dne and Ch is ophe B ooks, "E alua ing P edic i e Models o S uden Success: Closing he
Me hodological Gap," Me hodological Choices in Lea ning Analy ics, 2018. [Online]. A ailable: h ps://lea ning-
analy ics.in o/index.php/JLA/a icle/ iew/5814.
[4] S ephen Hu e al., "P ospec i ely p edic ing 4-yea college g adua ion om s uden applica ions," P oceedings
o he 8 h In e na ional Con e ence on Lea ning Analy ics and Knowledge, 2018. [Online]. A ailable:
h ps://dl.acm.o g/doi/10.1145/3170358.3170395.
[5] Kimbe ly E. A nold, Ma hew D. Pis illi, "Cou se signals a Pu due: using lea ning analy ics o inc ease s uden
success," P oceedings o he 2nd In e na ional Con e ence on Lea ning Analy ics and Knowledge, 2012. [Online].
A ailable: h ps://dl.acm.o g/doi/10.1145/2330601.2330666
[6] Ma k Da id Milli on, Lau a Malcolm and Da id Kil, "Insigh and Ac ion Analy ics: Th ee Case S udies o Conside ,"
RPajou nal, 2025. [Online]. A ailable: h ps://www. pajou nal.com/de /wp-con en /uploads/2014/10/A7.pd
[7] Ami Hossein Nabizadeh e al., "Lea ning Pa h Pe sonaliza ion and Recommenda ion Me hods: A Su ey o he
S a e-O -The-A ," Expe Sys ems wi h Applica ions, 2020. [Online]. A ailable:
h ps://www. esea chga e.ne /publica ion/341770404_Lea ning_Pa h_Pe sonaliza ion_and_Recommenda ion
_Me hods_A_Su ey_o _ he_S a e-O -The-A
[8] Ryan S. Bake e al., "Analyzing Ea ly A -Risk Fac o s in Highe Educa ion eLea ning Cou ses," P oceedings o he
8 h In e na ional Con e ence on Educa ional Da a Mining, 2014. [Online]. A ailable:
h ps:// iles.e ic.ed.go / ull ex /ED560553.pd
[9] Ca ie Klein e al., "Lea ning Analy ics Tools in Highe Educa ion: Adop ion a he In e sec ion o Ins i u ional
Commi men and Indi idual Ac ion," The Re iew o Highe Educa ion, 2019. [Online]. A ailable:
h ps://www. esea chga e.ne /publica ion/330069923_Lea ning_Analy ics_Tools_in_Highe _Educa ion_Adop i
on_a _ he_In e sec ion_o _Ins i u ional_Commi men _and_Indi idual_Ac ion.
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 3047-3053
3053
[10] René F. Kizilcec and Hansol Lee, "Algo i hmic Fai ness in Educa ion," a Xi :2007.05443 (cs), 2021. [Online].
A ailable: h p://a xi .o g/abs/2007.05443.
[11] Kelli Bi d, "P edic i e Analy ics in Highe Educa ion: The P omises and Challenges o Using Machine Lea ning o
Imp o e S uden Success," AIR P o essional File, 2023. [Online]. A ailable:
h ps://www. esea chga e.ne /publica ion/376840124_P edic i e_Analy ics_in_Highe _Educa ion_The_P omis
es_and_Challenges_o _Using_Machine_Lea ning_ o_Imp o e_S uden _Success
[12] Kimbe ly E. A nold, S e en Lonn and Ma hew D. Pis illi, "An exe cise in ins i u ional e lec ion: he lea ning
analy ics eadiness ins umen (LARI)," P oceedings o he Fou h In e na ional Con e ence on Lea ning
Analy ics And Knowledge, 2014. [Online]. A ailable: h ps://dl.acm.o g/doi/10.1145/2567574.2567621
