Co esponding au ho : Bola i o khadija A emu
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.
Ad anced machine lea ning-d i en business analy ics o eal- ime heal h isk
s a i ica ion and cos p edic ion models
Bola i o khadija A emu * and Olagunju Omo ola Peggy
Depa men o Applied Business Analy ics, Heal hca e T ack, Uni e si y o Wes Geo gia, USA.
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 150-167
Publica ion his o y: Recei ed on 23 Ma ch 2025; e ised on 28 Ap il 2025; accep ed on 01 May 2025
A icle DOI: h ps://doi.o g/10.30574/wja .2025.26.2.1583
Abs ac
The apid ad ancemen o machine lea ning (ML) echnologies has opened new on ie s in heal hca e analy ics,
pa icula ly in he domains o eal- ime heal h isk s a i ica ion and cos p edic ion modeling. T adi ional ac ua ial and
s a is ical me hods, while ounda ional, o en lack he agili y and p ecision necessa y o manage inc easingly complex
pa ien popula ions and dynamic heal h sys em demands. The in eg a ion o ad anced ML-d i en business analy ics
o e s a ans o ma i e pa hway, enabling heal hca e o ganiza ions o an icipa e clinical isks, alloca e esou ces mo e
e icien ly, and ansi ion owa d p oac i e, alue-based ca e deli e y models. This pape explo es he deploymen o
ad anced ML algo i hms—such as g adien boos ing, deep lea ning, and ensemble echniques— o p edic i e heal h
isk s a i ica ion a he indi idual and popula ion le els. I examines how eal- ime analy ics pla o ms, powe ed by
mul imodal pa ien da a and inancial indica o s, enable he con inuous e inemen o cos p edic ion models, leading
o mo e accu a e budge ing, a ge ed in e en ions, and isk-sha ing con ac s a egies. Emphasis is placed on he
c i ical ole o ea u e enginee ing, da a go e nance, model in e p e abili y, and e hical conside a ions, pa icula ly
a ound algo i hmic bias and anspa ency. D awing on case s udies om leading in eg a ed heal h sys ems and paye
o ganiza ions, he pape demons a es how p edic i e insigh s de i ed om ML models can educe hospi al admissions,
op imize ca e managemen p og ams, and imp o e inancial o ecas ing accu acy. By na owing om he b oade
con ex o ML's impac on heal hca e o speci ic applica ions in isk and cos p edic ion, his s udy p o ides ac ionable
amewo ks o in eg a ing ad anced analy ics in o s a egic heal hca e ope a ions and policymaking in an e a o ising
complexi y and accoun abili y.
Keywo ds: Machine Lea ning in Heal hca e; Real-Time Risk S a i ica ion; Cos P edic ion Models; P edic i e Business
Analy ics; Heal hca e Financial Fo ecas ing; Value-Based Ca e Analy ics
1. In oduc ion
1.1. Con ex : E olu ion o Analy ics in Heal hca e
The in eg a ion o da a analy ics in o heal hca e has unde gone a ema kable ans o ma ion o e he pas ew decades.
Ini ially, heal hca e analy ics ocused p ima ily on desc ip i e s a is ics de i ed om clinical and adminis a i e
da ase s, p o iding e ospec i e insigh s o unde s and disease pa e ns, pa ien lows, and ope a ional ine iciencies
[1]. Ea ly sys ems p ima ily elied on s uc u ed da a om elec onic heal h eco ds (EHRs) and insu ance claims,
p oducing basic me ics such as hospi al eadmission a es o a e age leng h o s ay [2].
As heal hca e deli e y g ew mo e complex, new demands eme ged o p edic i e and p esc ip i e analy ics capable o
p oac i ely managing pa ien isks, op imizing esou ce alloca ion, and imp o ing heal h ou comes. Inno a ions in
compu ing powe , da a s o age, and machine lea ning (ML) algo i hms enabled he shi om e ospec i e epo ing
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 150-167
151
o eal- ime and u u e-o ien ed analy ics [3]. Con empo a y heal h sys ems now collec as mul imodal da ase s
including imaging, genomic sequences, wea able senso ou pu s, and social de e minan s o heal h (SDOH), expanding
he analy ical ho izons a beyond adi ional clinical da a [4].
These echnological ad ancemen s, coupled wi h he global push owa d alue-based ca e models, ha e necessi a ed
mo e sophis ica ed analy ics amewo ks. P o ide s and paye s alike now seek o p edic ad e se e en s be o e hey
occu , pe sonalize ea men pa hways, and align ca e deli e y wi h inancial incen i es ied o ou comes a he han
olume [5]. The e o e, he e olu ion o heal hca e analy ics e lec s no only echnological ma u a ion bu also
undamen al changes in how heal h se ices a e inanced, egula ed, and deli e ed in he 21s cen u y [6].
1.2. Gap in T adi ional Me hods e sus New ML App oaches
Despi e no able p og ess, adi ional s a is ical me hods used in heal hca e analy ics ace signi ican limi a ions when
applied o complex, he e ogeneous pa ien popula ions. Classical isk s a i ica ion models, such as logis ic eg ession-
based sco ing sys ems, o en ely on a limi ed numbe o a iables, assume linea ela ionships, and a e no op imized
o de ec in ica e, non-linea pa e ns p esen in high-dimensional heal h da a [7]. Mo eo e , hey s uggle o
accommoda e eal- ime da a s eams and adap dynamically o e ol ing pa ien s a uses o ca e en i onmen s [8].
Con en ional app oaches a e also es ic ed in hei abili y o in eg a e mul imodal da a sou ces. Fo ins ance,
adi ional models may exclude i al insigh s om imaging da ase s, uns uc u ed clinical no es, o en i onmen al and
beha io al ac o s, leading o incomple e o biased isk assessmen s [9]. Fu he mo e, hese models ypically pe o m
s a ic p edic ions a a single poin in ime a he han con inuously upda ing isk p o iles as new da a becomes a ailable
[10].
Machine lea ning models o e a pa adigm shi by o e coming hese cons ain s. Algo i hms such as andom o es s,
g adien boos ing machines, and deep lea ning a chi ec u es can p ocess housands o ea u es simul aneously,
ecognize non-linea ela ionships, and imp o e p edic i e accu acy o e ime h ough i e a i e lea ning [11]. In
addi ion, ML models can be in eg a ed in o eal- ime clinical decision suppo sys ems, enabling dynamic and
pe sonalized in e en ions. As heal hca e sys ems pu sue mo e p oac i e and p e en i e ca e models, he limi a ions
o adi ional analy ics inc easingly highligh he necessi y o ad anced ML-d i en app oaches o manage popula ion
heal h isks and inancial ou comes e ec i ely [12].
1.3. Rele ance o Real-Time Risk S a i ica ion and Cos Fo ecas ing
Real- ime isk s a i ica ion and cos o ecas ing ep esen c i ical capabili ies o mode n heal hca e sys ems s i ing
o balance quali y imp o emen wi h cos con ainmen . Timely iden i ica ion o high- isk pa ien s allows p o ide s o
in e ene ea lie , p e en ing disease p og ession, a oiding unnecessa y hospi aliza ions, and educing he o e all
bu den on heal hca e in as uc u e [13]. Risk s a i ica ion models adi ionally p io i ized e ospec i e isk sco ing;
howe e , eal- ime sys ems dynamically inco po a e cu en i als, lab esul s, medica ion adhe ence, and o he
beha io al da a o upda e isk p o iles con inuously [14].
In pa allel, accu a e cos p edic ion models enable heal hca e adminis a o s o o ecas u u e expendi u es, alloca e
budge s mo e e ec i ely, and nego ia e alue-based con ac s g ounded in ealis ic isk assessmen s. T adi ional
ac ua ial me hods o en ail o cap u e he complexi y and a iabili y o heal hca e u iliza ion pa e ns, leading o ei he
unde es ima ion o o e es ima ion o inancial isks [15]. Ad anced machine lea ning models, le e aging mul imodal
da ase s, ha e demons a ed supe io pe o mance in p edic ing bo h indi idual pa ien cos s and popula ion-le el
inancial ends [16].
These capabili ies a e pa icula ly ele an in he con ex o global shi s owa d alue-based paymen models, such as
accoun able ca e o ganiza ions (ACOs) and bundled paymen ini ia i es, which penalize poo ou comes and ewa d cos
e iciency. By in eg a ing clinical, inancial, and ope a ional da a s eams, heal h sys ems can p oac i ely design
in e en ions, manage pa ien popula ions, and achie e sus ainable alue deli e y [17]. Consequen ly, eal- ime
analy ics is no longe a luxu y bu a s a egic impe a i e o heal hca e o ganiza ions seeking o h i e unde e ol ing
eimbu semen landscapes and ising pa ien expec a ions [18].
1.4. Objec i es o he Pape
This pape aims o explo e he in eg a ion o mul imodal pa ien da a and ad anced business in elligence (BI) sys ems
as pi o al enable s o s a egic heal hca e se ice op imiza ion and alue-based deli e y. Speci ically, i examines he
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 150-167
152
echnical, ope a ional, and e hical dimensions o le e aging machine lea ning-d i en analy ics o eal- ime heal h isk
s a i ica ion and cos o ecas ing [19].
Th ough a comp ehensi e e iew o cu en echnologies, sys em a chi ec u es, and case s udies, he pape seeks o:
• Analyze he e olu ion o heal hca e analy ics and he limi a ions o adi ional models.
• Desc ibe e ec i e s a egies o in eg a ing di e se da a modali ies in o uni ied analy ics pla o ms.
• Demons a e how BI ools can be employed o op imize esou ce alloca ion, enhance pa ien ca e, and suppo
alue-based paymen models.
• Iden i y challenges ela ed o da a go e nance, model in e p e abili y, and e hical isks.
• P opose u u e di ec ions o achie ing sus ainable, da a-d i en heal hca e ans o ma ion.
The indings aim o in o m heal hca e execu i es, clinicians, da a scien is s, and policymake s engaged in he digi al
ans o ma ion o heal hca e deli e y sys ems [20].
2. The e olu ion o heal hca e analy ics
2.1. T adi ional Heal hca e Risk S a i ica ion Models
T adi ional heal hca e isk s a i ica ion models ha e long been undamen al o popula ion heal h managemen ,
insu ance unde w i ing, and ca e coo dina ion ini ia i es. These models p ima ily elied on ac ua ial app oaches,
whe e la ge his o ical da ase s we e analyzed o gene a e isk sco es based on demog aphic ac o s, como bidi ies, and
p io heal hca e u iliza ion pa e ns [6]. Ac ua ial models ypically assumed homogenous isk wi hin s a a and
employed gene alized linea models o es ima e he p obabili y o ad e se e en s o high-cos u iliza ion. While hese
me hods p o ided a use ul s a ing poin o iden i ying a - isk indi iduals, hey o en lacked g anula i y and
adap abili y o di e se, dynamic pa ien popula ions [7].
Logis ic eg ession has his o ically been he co ne s one s a is ical me hod o heal hca e isk p edic ion. This app oach
models bina y ou comes, such as hospi al eadmission o mo ali y, based on a se o independen p edic o a iables
[8]. Logis ic models became he basis o nume ous clinical sco ing sys ems such as he Cha lson Como bidi y Index
(CCI), he LACE index o eadmission isk, and he APACHE sco e o in ensi e ca e uni mo ali y isk p edic ion [9].
These sco ing sys ems ope a ionalized isk s a i ica ion a he bedside, aiding clinical decision-making and esou ce
p io i iza ion.
Howe e , adi ional models ha e no able limi a ions. They o en depend on a limi ed numbe o a iables, equi e
manual a iable selec ion, and assume linea ela ionships be ween inpu s and ou comes. Mo eo e , hese models a e
gene ally s a ic—once buil , hey do no adap o incoming da a wi hou e aining—and a e ill-equipped o in eg a e
uns uc u ed da a sou ces such as physician no es, imaging s udies, o con inuous moni o ing da a [10]. As heal hca e
deli e y shi s owa d p ecision medicine and dynamic, eal- ime isk assessmen , hese adi ional me hods
inc easingly all sho in p o iding ac ionable insigh s o con empo a y heal hca e en i onmen s [11].
2.2. Rise o Machine Lea ning in Heal hca e Analy ics
The eme gence o machine lea ning (ML) has ca alyzed a undamen al ans o ma ion in heal hca e analy ics. Unlike
adi ional models, ML algo i hms can au onomously lea n pa e ns om la ge, complex da ase s wi hou being
explici ly p og ammed o ollow p ede e mined ela ionships. This lexibili y makes ML pa icula ly well-sui ed o
analyzing he mul i ace ed, high-dimensional da a inc easingly a ailable in heal hca e [12].
One o ML’s c i ical ad an ages is i s abili y o model non-linea and hie a chical ela ionships be ween a iables,
cap u ing in ica e in e ac ions ha adi ional me hods o en miss. Algo i hms such as andom o es s, suppo ec o
machines, and g adien boos ing machines can accommoda e hund eds o housands o a iables simul aneously,
au oma ically iden i ying he mos p edic i e ea u es wi hou human in e en ion [13]. Fu he mo e, deep lea ning
models, pa icula ly con olu ional and ecu en neu al ne wo ks, ha e demons a ed ema kable pe o mance in
p ocessing uns uc u ed da a ypes such as medical imaging, ee- ex clinical no es, and genomics [14].
Handling complex, high-dimensional da a has become inc easingly essen ial as heal hca e mo es beyond adi ional
EHR da a owa d in eg a ing mul imodal sou ces. Wea able senso s con inuously gene a e physiological signals; social
de e minan s o heal h da abases p o ide insigh s in o en i onmen al isks; and genomics in oduces as da ase s a
he molecula le el [15]. T adi ional s a is ical echniques o en al e when aced wi h such "wide" da ase s, whe e he
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 150-167
153
numbe o ea u es i als o exceeds he numbe o obse a ions. In con as , ML models excel a managing high ea u e-
o-sample a ios h ough echniques like egula iza ion, ea u e impo ance anking, and dimensionali y educ ion [16].
P edic i e accu acy is ano he domain whe e ML me hods consis en ly ou pe o m adi ional app oaches. S udies
compa ing ML models wi h logis ic eg ession in p edic ing hospi al eadmissions, sepsis onse , and mo ali y ha e
demons a ed supe io sensi i i y, speci ici y, and o e all a ea unde he ecei e ope a ing cha ac e is ic cu e
(AUROC) o ML app oaches [17]. Fo ins ance, he use o g adien boos ing algo i hms in p edic ing 30-day eadmission
a e hea ailu e hospi aliza ion has esul ed in 5–10% imp o emen s in p edic i e pe o mance o e adi ional
sco ing sys ems [18]. Mo eo e , ML models a e be e equipped o dynamically upda e isk sco es as new pa ien da a
becomes a ailable, o e ing eal- ime isk s a i ica ion capabili ies c i ical o p e en i e in e en ions and adap i e
ca e planning [19].
Howe e , he ise o ML in heal hca e analy ics is no wi hou challenges. Issues ela ed o model in e p e abili y,
gene alizabili y ac oss popula ions, and he isk o embedding sys emic biases in o algo i hms ha e spa ked signi ican
deba e among clinicians, da a scien is s, and policymake s [20]. Ne e heless, wi h app op ia e model go e nance
amewo ks, igo ous alida ion p ocesses, and e o s owa d explainable AI, ML-d i en heal hca e analy ics o e s
unp eceden ed oppo uni ies o imp o e pa ien ou comes, op imize esou ce use, and acili a e alue-based ca e
ans o ma ions.
Thus, he p og ession om adi ional ac ua ial and logis ic eg ession models o ad anced machine lea ning
ep esen s no me ely a echnological ad ancemen bu a pa adigm shi in how heal h isks a e concep ualized,
p edic ed, and ac ed upon in mode n heal hca e sys ems.
Figu e 1 Timeline o Heal hca e Analy ics E olu ion
3. Machine lea ning echniques o heal h isk s a i ica ion
3.1. Model T aining, Valida ion, and E alua ion
Rigo ous model aining, alida ion, and e alua ion p ocesses a e essen ial o de eloping obus machine lea ning
solu ions in heal hca e. P ope me hodology ensu es ha p edic i e models gene alize well ac oss di e se pa ien
popula ions and clinical se ings, a he han o e i ing o he idiosync asies o speci ic da ase s [21].
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 150-167
154
C oss- alida ion is a s anda d echnique o mi iga e o e i ing and assess model s abili y. K- old c oss- alida ion,
whe e he da a is spli in o k subse s and models a e ained and es ed i e a i ely, p o ides a mo e eliable es ima e o
model pe o mance compa ed o a single ain- es spli . S a i ied k- old c oss- alida ion is pa icula ly impo an in
heal hca e scena ios in ol ing class imbalance, ensu ing p opo ional ep esen a ion o a e ou comes like mo ali y
o ad e se d ug e en s in each old [22].
Hype pa ame e uning u he op imizes model pe o mance. G id sea ch, andom sea ch, and Bayesian op imiza ion
me hods sys ema ically explo e he hype pa ame e space o iden i y se ings ha maximize accu acy, minimize loss,
o o he wise mee speci ic pe o mance c i e ia. In ensemble me hods such as andom o es s and GBMs, uning
hype pa ame e s like ee dep h, lea ning a e, and he numbe o es ima o s can signi ican ly enhance esul s [23].
Once ained, models a e e alua ed using mul iple pe o mance me ics o cap u e hei p edic i e u ili y
comp ehensi ely. The a ea unde he ecei e ope a ing cha ac e is ic cu e (AUROC) is widely used o measu e
disc imina ion abili y; a model wi h an AUROC close o 1.0 indica es excellen sepa abili y be ween classes [24].
Howe e , AUROC may no adequa ely e lec pe o mance in imbalanced da ase s, whe e mino i y class de ec ion is
pa amoun .
In such cases, p ecision- ecall (PR) cu es p o ide a mo e sensi i e e alua ion. P ecision (posi i e p edic i e alue)
quan i ies how many p edic ed posi i es a e ue posi i es, while ecall (sensi i i y) assesses he p opo ion o ac ual
posi i es co ec ly iden i ied. The F1 sco e, he ha monic mean o p ecision and ecall, balances hese me ics and is
pa icula ly use ul o op imizing models whe e bo h alse posi i es and alse nega i es ca y signi ican clinical isks
[25].
Con usion ma ices, calib a ion plo s, and decision cu e analysis add u he laye s o alida ion, illus a ing model
eliabili y ac oss a ious p obabili y h esholds. Mo eo e , model in e p e abili y echniques—such as SHAP (SHapley
Addi i e exPlana ions) alues and ea u e impo ance ankings—a e inc easingly in eg a ed in o e alua ion p o ocols
o ensu e clinicians and adminis a o s can us and ac on model p edic ions [26].
Ul ima ely, success ul deploymen o ML models o heal h isk s a i ica ion hinges no solely on high p edic i e me ics
bu also on main aining anspa ency, ai ness, and clinical ele ance h oughou he de elopmen li ecycle.
Table 1 Compa ison o Key Machine Lea ning Algo i hms o Risk S a i ica ion
Algo i hm
Type
Ad an ages
Limi a ions
Typical
Heal hca e Use
Cases
Decision T ee
Supe ised Lea ning
(Classi ica ion/Reg ession)
Easy o in e p e ,
as o ain
P one o
o e i ing, low
p edic i e powe
alone
Basic isk sco ing,
mo ali y
p edic ion
Random Fo es
Ensemble (Bagging)
Handles non-linea
da a, educes
o e i ing
Less
in e p e able,
slowe in e ence
Readmission
p edic ion,
ch onic disease
managemen
G adien Boos ing
Machine (GBM,
XGBoos ,
Ligh GBM)
Ensemble (Boos ing)
High p edic i e
accu acy, handles
missing da a
P one o
o e i ing i no
uned, slowe
aining
Sepsis p edic ion,
ea ly
de e io a ion
de ec ion
Con olu ional
Neu al Ne wo k
(CNN)
Deep Lea ning
Excels in image and
spa ial da a
analysis
Requi es la ge
labeled da ase s,
complex o ain
Radiology image
classi ica ion,
umo de ec ion
Recu en Neu al
Ne wo k (RNN)
and LSTM
Deep Lea ning
E ec i e o
sequen ial/ ime
se ies da a
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 150-167
155
Figu e 2 Pipeline o ML Model De elopmen in Heal hca e
4. Real- ime da a in eg a ion and business in elligence pla o ms
4.1. Sou ces o Mul imodal Da a o Risk P edic ion
Risk p edic ion in heal hca e has e ol ed signi ican ly h ough he in eg a ion o mul imodal da a sou ces ha ex end
beyond adi ional clinical eco ds. The p ima y ounda ion emains elec onic heal h eco ds (EHRs), which p o ide
s uc u ed clinical in o ma ion such as diagnoses, labo a o y esul s, medica ion his o ies, and p ocedu es. EHRs enable
longi udinal acking o pa ien in e ac ions ac oss di e en ca e se ings, o ming a c ucial da ase o p edic i e
modeling [15].
Claims da a, gene a ed om billing and insu ance p ocesses, o e a complemen a y iew o heal hca e u iliza ion
pa e ns, including equency o admissions, specialis e e als, and medica ion adhe ence. Al hough some imes
lacking clinical g anula i y, claims da ase s a e in aluable o iden i ying high-cos u ilize s and es ima ing o al cos o
ca e [16].
Wea able de ices in oduce con inuous s eams o eal- ime biome ic da a, including hea a e, blood p essu e,
glucose le els, and physical ac i i y pa e ns. These da a s eams p o ide dynamic physiological ma ke s ha
adi ional da ase s may miss, allowing ea lie iden i ica ion o heal h de e io a ion o li es yle- ela ed isks [17]. The
adop ion o wea able echnology in ch onic disease managemen , pa icula ly in ca diology and diabe es ca e,
unde sco es i s g owing alue o p edic i e analy ics.
Social de e minan s o heal h (SDOH) ep esen ano he c i ical ye his o ically unde u ilized da a domain. In o ma ion
abou a pa ien ’s educa ion le el, income, housing s abili y, and access o anspo a ion signi ican ly in luences heal h
ou comes and heal hca e cos s. Inco po a ing SDOH me ics in o isk models en iches p edic i e pe o mance and
suppo s mo e equi able esou ce alloca ion [18].
By me ging clinical, inancial, beha io al, and socio-en i onmen al da ase s, mode n heal h sys ems can cons uc mo e
comp ehensi e, pe sonalized, and con ex -awa e isk p o iles, mo ing close o uly p oac i e heal hca e deli e y.
4.2. Real-Time Da a S eaming and P ocessing
E icien eal- ime da a s eaming and p ocessing a chi ec u es a e essen ial o le e aging mul imodal heal hca e da a
in o ac ionable insigh s. T adi ional ba ch-p ocessing sys ems, whe e da a a e collec ed and analyzed a e signi ican
delays, a e inadequa e o applica ions equi ing immedia e clinical decision suppo [19].
Applica ion p og amming in e aces (APIs) ha e become he s anda d me hod o enabling in e ope abili y be ween
dispa a e heal h in o ma ion sys ems. APIs acili a e he secu e, on-demand exchange o da a be ween EHRs, labo a o y
sys ems, wea able de ices, and analy ics pla o ms. S anda ds-based APIs, such as hose ollowing he Heal h Le el
Se en (HL7) Fas Heal hca e In e ope abili y Resou ces (FHIR) p o ocol, a e inc easingly adop ed o enhance
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 150-167
156
compa ibili y and simpli y in eg a ion e o s [20]. FHIR APIs allow o modula access o speci ic heal hca e esou ces
(e.g., pa ien demog aphics, medica ion lis s), enabling de elope s o build esponsi e, pa ien -cen e ed applica ions.
Beyond APIs, eal- ime da a a chi ec u es equi e obus s eaming echnologies. Apache Ka ka, a dis ibu ed e en
s eaming pla o m, has eme ged as a leading solu ion o inges ing and p ocessing la ge olumes o eal- ime
heal hca e da a. Ka ka enables he con inuous mo emen o e en da a—such as pa ien i als om emo e moni o ing
de ices o medica ion adminis a ion eco ds—in o analy ics engines wi h minimal la ency [21].
Cloud-based s eaming se ices such as Amazon Web Se ices (AWS) Kinesis o e scalable al e na i es, pa icula ly o
o ganiza ions seeking o minimize on-p emises in as uc u e cos s. Kinesis acili a es he inges ion, p ocessing, and
analysis o s eaming heal hca e da a wi hin milliseconds, suppo ing use cases like ea ly wa ning sys ems o sepsis o
eal- ime bed managemen dashboa ds [22].
By le e aging APIs, HL7/FHIR s anda ds, and scalable s eaming a chi ec u es, heal hca e o ganiza ions can es ablish
eal- ime analy ics ecosys ems ha in eg a e di e se da a sou ces and deli e insigh s apidly enough o impac clinical
decision-making and ope a ional e iciency.
4.3. Business In elligence In eg a ion o Decision Suppo
While he inges ion and p ocessing o mul imodal heal hca e da a is c ucial, he eal alue eme ges when insigh s a e
e ec i ely in eg a ed in o on line decision suppo h ough business in elligence (BI) pla o ms. BI ools ans o m
aw da a in o ac ionable in o ma ion ia dashboa ds, p edic i e ale s, and wo k low-in eg a ed ecommenda ions ha
can be accessed by clinicians, adminis a o s, and ca e manage s [23].
Real- ime dashboa ds p o ide an in ui i e in e ace o isualizing key heal h sys em me ics. In he con ex o isk
s a i ica ion, dashboa ds may display isk sco es o indi idual pa ien s, highligh ends in hospi al admissions, o
o ecas bed occupancy le els. E ec i e dashboa ds balance simplici y and dep h, o e ing d ill-down capabili ies ha
allow use s o in es iga e anomalies wi hou o e whelming hem wi h unnecessa y complexi y [24]. Mo eo e ,
dashboa ds ha a e embedded wi hin exis ing clinical sys ems (e.g., EHRs) p omo e adop ion by educing wo k low
dis up ions and cogni i e bu den on use s.
P edic i e ale s ep esen ano he c i ical applica ion o BI in eg a ion. Fo example, an ea ly wa ning sys em
embedded wi hin an EHR can ale ca e eams when a pa ien 's i als c oss a h eshold indica i e o impending clinical
de e io a ion. P edic i e ale s mus be ca e ully uned o balance sensi i i y and speci ici y, minimizing he isk o ale
a igue ha can desensi ize clinicians o impo an wa nings [25]. Bes p ac ices include isk s a i ica ion h esholds
adjus able by clinicians, ie ed ale sys ems based on u gency, and in eg a ion wi h escala ion p o ocols o ensu e
imely in e en ions.
In eg a ion wi h clinical wo k lows is pe haps he mos c i ical success ac o . S andalone analy ics pla o ms
disconnec ed om daily clinical ou ines o en ail o impac ca e deli e y meaning ully. Embedding p edic i e models
wi hin clinical pa hways—such as au oma ically sugges ing ca e managemen en ollmen o high- isk pa ien s du ing
ou pa ien isi s—ensu es ha insigh s a e ope a ionalized a he poin o ca e [26]. Fu he mo e, closed-loop analy ics
amewo ks ha ack he ou comes o p edic i e in e en ions p o ide con inuous eedback o e ine and ecalib a e
models o e ime.
E ec i e BI in eg a ion also equi es go e nance s uc u es ha p omo e anspa ency, accoun abili y, and clinician
engagemen . Model alida ion epo s, pe o mance dashboa ds showing p edic ion accu acy, and clinician educa ion
ini ia i es help build us and encou age esponsible adop ion o p edic i e analy ics ools [27].
Ul ima ely, he seamless in eg a ion o eal- ime da a s eams in o BI pla o ms capable o suppo ing clinical and
ope a ional decisions is wha ans o ms mul imodal heal hca e da a om a passi e eposi o y in o an ac i e agen o
heal hca e se ice op imiza ion, be e pa ien ou comes, and sus ainable alue-based deli e y.
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 150-167
157
Figu e 3 A chi ec u e o Real-Time Heal h Risk s a i ica ion sys em
5. Cos p edic ion models powe ed by machine lea ning
5.1. T adi ional Cos P edic ion App oaches
Cos p edic ion has long been a cen al objec i e o heal hca e analy ics, pa icula ly o insu e s, go e nmen agencies,
and hospi al inancial o ice s. T adi ional app oaches elied hea ily on s a is ical eg ession models and ac ua ial isk
adjus men echniques o o ecas heal hca e expendi u es a bo h indi idual and popula ion le els [19].
Linea eg ession models we e among he ea lies me hods used, p edic ing u u e cos s based on a ela i ely small se
o a iables such as age, gende , diagnosis his o y, and p io -yea expenses. These models assumed linea i y,
independence, and homoscedas ici y—assump ions o en iola ed in complex heal hca e da ase s cha ac e ized by
skewed dis ibu ions and he e ogeneous pa ien beha io s [20].
Mo e sophis ica ed gene alized linea models (GLMs), including Poisson and nega i e binomial eg ession, we e
in oduced o accommoda e he o e dispe sion common in heal hca e cos da a. Despi e imp o emen s, GLMs s ill
s uggled o cap u e non-linea ela ionships o in e ac ions among isk ac o s, limi ing hei p edic i e accu acy [21].
Ac ua ial isk adjus men me hods, such as he Hie a chical Condi ion Ca ego y (HCC) model used in Medica e
Ad an age p og ams, adjus ed paymen s o insu e s based on en ollees’ isk p o iles. These models agg ega e
como bidi y g oupings o gene a e isk sco es ha p edic heal hca e cos s ela i e o a popula ion a e age. While
ac ua ial models ha e se ed well o policy pu poses, hey a e ypically calib a ed annually, ail o inco po a e eal-
ime da a, and unde pe o m when p edic ing cos s a an indi idual le el [22].
O e all, adi ional app oaches p o ided aluable bu coa se-g ained o ecas s sui able o la ge coho s bu
insu icien ly g anula o dynamic o p ecision budge ing, a ge ed in e en ions, o eal- ime inancial planning in
e ol ing heal hca e en i onmen s [23].
5.2. Ad anced ML Techniques o Cos Fo ecas ing
Ad anced machine lea ning (ML) echniques ha e e olu ionized heal hca e cos p edic ion by add essing he
limi a ions o adi ional s a is ical models. Ensemble models such as andom o es s, g adien boos ing machines
(GBMs), and XGBoos ha e shown supe io pe o mance in modeling he non-linea , high-dimensional ela ionships
inhe en in heal hca e spending beha io s [24].
Random o es s, by agg ega ing p edic ions om mul iple decision ees, educe o e i ing and imp o e
gene alizabili y, making hem ideal o cap u ing complex in e ac ions among socio-demog aphic, clinical, and
beha io al ea u es a ec ing heal hca e cos s. GBMs sequen ially co ec he e o s o p e ious models, enhancing
p edic i e p ecision, especially in da ase s wi h a e bu high-cos e en s [25].
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 150-167
158
Deep lea ning a chi ec u es u he expand he on ie o cos o ecas ing. A i icial neu al ne wo ks (ANNs) and deep
neu al ne wo ks (DNNs) can inges as amoun s o s uc u ed and uns uc u ed da a, disco e ing la en pa e ns
wi hou equi ing explici ea u e enginee ing. In ecen s udies, deep lea ning models inco po a ing EHR da a, claims
his o y, and social de e minan s ha e ou pe o med adi ional GLMs and e en classical ML models in o ecas ing one-
yea heal hca e expendi u es [26].
Mo eo e , ecu en neu al ne wo ks (RNNs) and hei a ian s, pa icula ly long sho - e m memo y (LSTM)
ne wo ks, excel a handling empo al sequences. These a chi ec u es a e well-sui ed o p edic ing cos s based on
longi udinal pa ien his o ies, dynamically adjus ing isk p o iles o e ime a he han elying on s a ic snapsho s [27].
Impo an ly, ad anced ML models can con inuously lea n om incoming da a, allowing o eal- ime o nea - eal- ime
upda ing o cos o ecas s. This dynamic adap abili y is c i ical o mode n heal h sys ems ope a ing unde accoun able
ca e con ac s and bundled paymen a angemen s whe e inancial and clinical isks mus be cons an ly moni o ed and
mi iga ed [28].
5.3. Impac o Accu a e Cos P edic ion on Heal h Sys ems
Accu a e cos p edic ion models ha e a ans o ma i e impac on heal hca e o ganiza ions, enabling mo e p ecise
budge ing, esou ce alloca ion, and con ac nego ia ion unde alue-based ca e amewo ks. Reliable cos o ecas ing
allows hospi al adminis a o s o an icipa e inancial needs, alloca e s a and in as uc u e e icien ly, and p e en cos
o e uns ha h ea en ope a ional sus ainabili y [29].
Budge ing p ocesses adi ionally depended on his o ical a e ages o b oad ac ua ial adjus men s, which o en ailed
o accoun o he complexi y o pa ien popula ions o e ol ing ea men pa adigms. ML-d i en cos models can
p ojec cos s a indi idual, depa men al, and o ganiza ional le els wi h a g ea e speci ici y, suppo ing g anula
inancial planning and capi al in es men decisions [30].
In e ms o esou ce alloca ion, p edic i e cos analy ics iden i y high- isk, high-cos pa ien coho s ea ly, allowing
a ge ed in e en ions such as in ensi e ca e managemen p og ams, eleheal h moni o ing, o social suppo se ices.
These in e en ions no only imp o e pa ien ou comes bu also lowe u u e cos s by p e en ing a oidable
hospi aliza ions, eme gency isi s, and complica ions [31].
Unde alue-based con ac ing models—such as sha ed sa ings ag eemen s, bundled paymen s, and capi a ion
a angemen s—accu a e cos p edic ion is essen ial o managing inancial isk. P o ide s en e ing hese con ac s
mus es ima e expec ed expendi u es accu a ely o se app op ia e isk co ido s and bonus s uc u es. ML-enhanced
o ecas ing empowe s o ganiza ions o nego ia e con ac s om a posi ion o da a-d i en con idence a he han
specula i e isk [32].
Mo eo e , p edic i e cos modeling suppo s popula ion heal h managemen ini ia i es by in o ming he design o ca e
pa hways ailo ed o he inancial and clinical isk p o iles o di e en g oups. When in eg a ed in o business
in elligence dashboa ds, hese o ecas s o e eal- ime insigh s o execu i e eams, inance depa men s, and clinical
leade s, aligning ope a ional and s a egic decision-making wi h o e a ching alue-based ca e goals [33].
Ul ima ely, imp o ed cos o ecas ing capabili ies d i en by ad anced analy ics se e no only o p o ec heal hca e
o ganiza ions inancially bu also o enhance hei abili y o deli e e icien , high-quali y, and equi able ca e in an
inc easingly complex and compe i i e landscape.
Table 2 Pe o mance Compa ison Be ween T adi ional and ML Cos P edic ion Models
Model
Mean Absolu e
E o (MAE)
Roo Mean
Squa e E o
(RMSE)
Mean Absolu e
Pe cen age E o
(MAPE)
No es
Linea Reg ession
4,800
6,200
18%
Baseline model; limi ed
handling o non-linea i ies
Gene alized Linea
Model (GLM)
4,500
5,850
16%
Be e o skewed cos da a
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 150-167
165
ollowing HL7 FHIR s anda ds, and eal- ime s eaming solu ions like Ka ka o AWS Kinesis, es ablish he echnological
backbone o dynamic da a inges ion and p ocessing.
On his ounda ion, p edic i e models a e buil , igo ously ained, alida ed, and moni o ed using bes p ac ices ha
emphasize c oss- alida ion, hype pa ame e uning, and bias mi iga ion. Fea u e enginee ing emains a c i ical s ep,
ensu ing ha models cap u e meaning ul clinical and beha io al signals a he han spu ious co ela ions.
Once models a e alida ed, hei ou pu s mus be ope a ionalized h ough eal- ime business in elligence pla o ms.
Risk sco es and cos o ecas s should be embedded di ec ly in o elec onic heal h eco d sys ems, adminis a i e
dashboa ds, and ca e managemen wo k lows o ensu e ha insigh s d i e imely and ac ionable in e en ions.
P edic i e ale s, dynamic esou ce alloca ion ecommenda ions, and pe sonalized ca e pa hway sugges ions mus be
in ui i ely accessible o on line use s wi hou in oducing wo k low dis up ions.
Con inuous eedback loops a e essen ial o main aining model pe o mance o e ime. Real-wo ld ou comes om
p edic i e in e en ions should be cap u ed and used o ecalib a e models, ensu ing ha hey emain aligned wi h
e ol ing clinical eali ies and o ganiza ional p io i ies. Thus, success ul in eg a ion o machine lea ning in o heal hca e
ope a ions is a sys emic e o equi ing ha moniza ion o echnology, p ocesses, go e nance s uc u es, and human
ac o s.
9.3. Final Re lec ions on E hical, Ope a ional, and Technological P io i ies o Fu u e Adop ion
As heal hca e con inues i s digi al ans o ma ion, he u u e success o machine lea ning-d i en analy ics will depend
hea ily on add essing a io o in e wined p io i ies: e hical in eg i y, ope a ional obus ness, and echnological
inno a ion.
E hically, heal h sys ems mus emain s ead as in p o ec ing pa ien p i acy, ensu ing da a secu i y, and p omo ing
ai ness in p edic i e modeling. The ad en o ede a ed lea ning, di e en ial p i acy echniques, and explainable AI
amewo ks o e s powe ul ools o balance he hi s o analy ic insigh s wi h he impe a i es o indi idual igh s and
socie al equi y. None heless, embedding e hical igilance in o e e y s age o he machine lea ning li ecycle— om da a
collec ion o deploymen —is a non-nego iable esponsibili y.
Ope a ionally, o ganiza ions mus in es no only in echnology bu also in cul u e. Building in e disciplina y eams ha
combine clinical, analy ical, IT, and go e nance expe ise is i al o designing and sus aining impac ul ML ini ia i es.
Clinician engagemen , anspa en alida ion p o ocols, and pa icipa o y design p ocesses can signi ican ly enhance
model adop ion, ensu ing ha p edic i e ools complemen a he han complica e heal hca e deli e y. Fu he mo e,
heal hca e leade s mus champion analy ics li e acy among s a , os e ing an en i onmen whe e da a-d i en decision-
making becomes he no m a he han he excep ion.
Technologically, he u u e demands con inuous inno a ion. Models mus e ol e o in eg a e new da a s eams such as
genomic p o iles, en i onmen al exposu es, and pa ien - epo ed ou comes. Edge compu ing, 5G-enabled emo e
moni o ing, and eal- ime analy ics a he poin o ca e will inc easingly shape he analy ics landscape. Addi ionally,
dynamic model upda ing mechanisms, au oma ed bias de ec ion sys ems, and human-in- he-loop AI go e nance
amewo ks will be c i ical o main aining model ele ance and us wo hiness o e ime.
In conclusion, machine lea ning holds unp eceden ed p omise o op imizing heal hca e se ices and achie ing ue
alue-based ca e. Realizing his p omise, howe e , will equi e unwa e ing commi men o e hical s ewa dship,
ope a ional excellence, and echnological agili y.
Compliance wi h e hical s anda ds
Disclosu e o con lic o in e es
No con lic o in e es o be disclosed.
Re e ences
[1] Rehan H. A i icial in elligence and machine lea ning: The impac o machine lea ning on p edic i e analy ics in
heal hca e. Inno a i e Compu e Sciences Jou nal. 2023 Ma 17;9(1):1-20.
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 150-167
166
[2] Abe R. U ilizing P edic i e Insigh s o Fu u e Planning: Rede ining Choices wi h Ad anced Da a Solu ions.
[3] Noah GU. In e disciplina y s a egies o in eg a ing o al heal h in na ional immune and in lamma o y disease
con ol p og ams. In J Compu Appl Technol Res. 2022;11(12):483-498. doi:10.7753/IJCATR1112.1016.
[4] Singh B, Nayya A. Na iga ing deep lea ning models and heal h moni o ing in as uc u e inancing in sma
ci ies: Re iew om legal pe cep ions and u u e inno a ions. Deep Lea ning in Enginee ing, Ene gy and Finance.
2024 Dec 26:80-114.
[5] S ini asagopalan LN. P edic i e Analy ics Fo Heal hca e Insu ance Risk Assessmen Using Ensemble Lea ning
Models. Jou nal ID. 2024 Jun 22;4195:6829.
[6] Odumbo O, Aso ose E, Oluwagbade E, Alemede V. Reenginee ing sus ainable pha maceu ical supply chains o
imp o e he apeu ic equi y in U.S. unde se ed heal h egions. In J Eng Technol Res Manag. 2024 Jun;8(6):208.
A ailable om: h ps://doi.o g/10.5281/zenodo.15289162
[7] Gudala M, Koilakonda RR. The Impac o A i icial In elligence and Machine Lea ning on Business Analy ics.
In e na ional Resea ch Jou nal o Economics and Managemen S udies IRJEMS. 2024;3(8).
[8] Okeke CMG. E alua ing company pe o mance: he ole o EBITDA as a key inancial me ic. In J Compu Appl
Technol Res. 2020;9(12):336–349
[9] Chukwunweike Joseph, Salaudeen Habeeb Dolapo. Ad anced Compu a ional Me hods o Op imizing Mechanical
Sys ems in Mode n Enginee ing Managemen P ac ices. In e na ional Jou nal o Resea ch Publica ion and
Re iews. 2025 Ma ;6(3):8533-8548. A ailable om: h ps://ij p .com/uploads/V6ISSUE3/IJRPR40901.pd
[10] Mish a R, T ipa hi P, Kuma N. Fu u e Di ec ions in he Applica ion o Machine Lea ning and In elligen
Op imiza ion in Business Analy ics. InIn elligen Op imiza ion Techniques o Business Analy ics 2024 (pp. 49-
76). IGI Global.
[11] An hony OC, Oluwagbade E, Baka e A, Animasahun B. E alua ing he economic and clinical impac s o
pha maceu ical supply chain cen aliza ion h ough AI-d i en p edic i e analy ics: compa a i e lessons om
la ge-scale cen alized p ocu emen sys ems and implica ions o d ug p icing, a ailabili y, and ca dio ascula
heal h ou comes in he U.S. In J Res Publ Re . 2024 Oc ;5(10):5148-5161. A ailable om:
h ps://ij p .com/uploads/V5ISSUE10/IJRPR34458.pd
[12] Adekunle BI, Chukwuma-Eke EC, Balogun ED, Ogunsola KO. A p edic i e modeling app oach o op imizing
business ope a ions: A case s udy on educing ope a ional ine iciencies h ough machine lea ning. In e na ional
Jou nal o Mul idisciplina y Resea ch and G ow h E alua ion. 2021;2(1):791-9.
[13] Alha be N, Almalki M. IoT-enabled heal hca e ans o ma ion le e aging deep lea ning o ad anced pa ien
moni o ing and diagnosis. Mul imedia Tools and Applica ions. 2024 Jul 30:1-4.
[14] Ag awal R, Desai S, Dholwani D, Keda i N, Bane jee A. A i icial In elligence/Machine Lea ning d i en decision
making in business analy ics o Financial Sec o using Ensemble Machine Lea ning Techniques. In2024 IEEE 3 d
Wo ld Con e ence on Applied In elligence and Compu ing (AIC) 2024 Jul 27 (pp. 310-315). IEEE.
[15] Emi-Johnson Oluwabukola, Fasanya Oluwa unmibi, Adeniyi Ayodele. P edic i e c op p o ec ion using machine
lea ning: A scalable amewo k o U.S. Ag icul u e. In J Sci Res A ch. 2024;15(01):670-688. A ailable om:
h ps://doi.o g/10.30574/ijs a.2024.12.2.1536
[16] Mu ugan MS. La ge-scale da a-d i en inancial isk managemen & analysis using machine lea ning s a egies.
Measu emen : Senso s. 2023 Jun 1;27:100756.
[17] Oluwagbade E, Animasahun B, Baka e A, An hony OC. Le e aging machine lea ning o p edic medica ion
adhe ence and ca dio ascula ou comes unde cen alized pha maceu ical p ocu emen : a Medica e Pa D
analysis. In J Eng Technol Res Manag. 2024; A ailable om: h ps://doi.o g/10.5281/zenodo.15194544
[18] Singh B, Nayya A. Na iga ing deep lea ning. Deep Lea ning in Enginee ing, Ene gy and Finance: P incipals and
Applica ions. 2024 Dec 26:80.
[19] Olagunju E. In eg a ing AI-d i en demand o ecas ing wi h cos -e iciency models in biopha maceu ical
dis ibu ion sys ems. In J Eng Technol Res Manag [In e ne ]. 2022 Jun 6(6):189. A ailable om:
h ps://doi.o g/10.5281/zenodo.15244666
[20] Fa ima S. Imp o ing Heal hca e Ou comes h ough Machine Lea ning: Applica ions and Challenges in Big Da a
Analy ics. In e na ional Jou nal o Ad anced Resea ch in Enginee ing Technology & Science. 2024;11.
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 150-167
167
[21] Odeni an O M (Janua y 30, 2025) Explo ing he Po en ial o Bamba a G oundnu Flou as an Al e na i e o
Diabe ic and Obese Pa ien s in he USA: A Comp ehensi e Re iew. Cu eus 17(1): e78258.
doi:10.7759/cu eus.78258
[22] Machi eddy JR. In eg a ing Machine Lea ning-D i en RPA wi h Cloud-Based Da a Wa ehousing o Real-Time
Analy ics and Business In elligence. Hong Kong Jou nal o AI and Medicine. 2024 May 2;4(1):98-121.
[23] Emi-Johnson Oluwabukola, Nk umah Kwame, Folasole Ade ayo, Amusa Tope Kolade. Op imizing machine
lea ning o imbalanced classi ica ion: Applica ions in U.S. heal hca e, inance, and secu i y. In J Eng Technol Res
Manag. 2023 No ;7(11):89. A ailable om: h ps://doi.o g/10.5281/zenodo.15188490
[24] Ke hu SS, Na la S, Vali a hi DT, Peddi S, Na a ajan DR. Pa ien -cen ic machine lea ning me hods and AI ools
o p edic ing and managing ch onic condi ions in elde ly ca e: Algo i hmic insigh s om he SURGE-Ahead
P ojec . ISAR-In . J. Res. Eng. Technol. 2023;8(1):28.
[25] Olagunju E. In eg a ing AI-d i en demand o ecas ing wi h cos -e iciency models in biopha maceu ical
dis ibu ion sys ems. In J Eng Technol Res Manag [In e ne ]. 2022 Jun 6(6):189. A ailable om:
h ps://doi.o g/10.5281/zenodo.15244666
[26] Roy PD, Chowdho y UG, Dey A, Sago DH. AI and Machine Lea ning in Heal hca e: Ad ancing Diagnos ics,
Pe sonalized T ea men , and P edic i e Modeling.
[27] Chukwunweike J, Lawal OA, A ogundade JB, Alade B. Na iga ing e hical challenges o explainable AI in
au onomous sys ems. In e na ional Jou nal o Science and Resea ch A chi e. 2024;13(1):1807–19.
doi:10.30574/ijs a.2024.13.1.1872. A ailable om: h ps://doi.o g/10.30574/ijs a.2024.13.1.1872.
[28] Kehinde O. Le e aging Da a-D i en Decision-Making o Enhanced Risk Managemen and Resou ce Alloca ion in
P ojec s.
[29] Oluwagbade E, Alemede V, Odumbo O, Animasahun B. Li ecycle go e nance o explainable AI in pha maceu ical
supply chains: a amewo k o con inuous alida ion, bias audi ing, and equi able heal hca e deli e y. In J Eng
Technol Res Manag. 2023 No ;7(11):54. A ailable om: h ps://doi.o g/10.5281/zenodo.15124514
[30] Sa ke M. Re olu ionizing heal hca e: he ole o machine lea ning in he heal h sec o . Jou nal o A i icial
In elligence Gene al science (JAIGS) ISSN: 3006-4023. 2024 Feb 27;2(1):36-61.
[31] The hi SK. Machine lea ning models o cos -e ec i e heal hca e deli e y sys ems: A global pe spec i e. Digi al
T ans o ma ion in Heal hca e. 2024 May 6;5:199.
[32] Kababii o Lillian. Ha nessing A i icial In elligence o Real-Time Compliance in he U.S. Oil & Gas Sec o :
Enhancing Tax Accu acy, Cu bing E asion, and Unlocking Re enue G ow h h ough In elligen Au oma ion.
In e na ional Jou nal o Compu e Applica ions Technology and Resea ch. 2025;14(05):55–70.
doi:10.7753/IJCATR1405.1006.
[33] Joseph Nnaemeka Chukwunweike and Opeyemi A o. Implemen ing agile managemen p ac ices in he e a o
digi al ans o ma ion [In e ne ]. Vol. 24, Wo ld Jou nal o Ad anced Resea ch and Re iews. GSC Online P ess;
2024. A ailable om: DOI: 10.30574/wja .2024.24.1.3253
[34] Ponnusamy V, Nalla asan V, Rajasega RS, A i azhagan N, Gou haman P. Mode n eal-wo ld applica ions using
da a analy ics and machine lea ning. InDa a Analy ics and Machine Lea ning: Na iga ing he Big Da a Landscape
2024 Ma 20 (pp. 215-235). Singapo e: Sp inge Na u e Singapo e.
[35] S ielkowski W, Vlaso A, Seli ano K, Mu a ie K, Shakhno V. P ospec s and challenges o he machine lea ning
and da a-d i en me hods o he p edic i e analysis o powe sys ems: A e iew. Ene gies. 2023 May
11;16(10):4025.
[36] Ba i MD, A a A. The Impac O Machine Lea ning On P esc ip i e Analy ics Fo Op imized Business Decision-
Making. Anjuman, The Impac O Machine Lea ning On P esc ip i e Analy ics Fo Op imized Business Decision-
Making (Ap il 15, 2024). 2024 Ap 15.
[37] Osman A, Fowowe OO, Agboluaje R, O ekha PO. In eg a ing machine lea ning in business analy ics consul ing o
p oac i e decision-making and inno a ion.
[38] Olayinka OH. Da a d i en cus ome segmen a ion and pe sonaliza ion s a egies in mode n business in elligence
amewo ks. Wo ld Jou nal o Ad anced Resea ch and Re iews. 2021;12(3):711–726. doi:
h ps://doi.o g/10.30574/wja .2021.12.3.0658.
