Machine Learning First Response to COVID-19: A Systematic Literature Review of Clinical Decision Assistance Approaches during Pandemic Years from 2020 to 2022

Ci a ion: Badiola-Zabala, G.;
Lopez-Guede, J.M.; Es e ez, J.; G aña,
M. Machine Lea ning Fi s Response
o COVID-19: A Sys ema ic Li e a u e
Re iew o Clinical Decision
Assis ance App oaches du ing
Pandemic Yea s om 2020 o 2022.
Elec onics 2024,13, 1005. h ps://
doi.o g/10.3390/elec onics13061005
Academic Edi o : Gab iella Olmo
Recei ed: 15 Janua y 2024
Re ised: 27 Feb ua y 2024
Accep ed: 3 Ma ch 2024
Published: 7 Ma ch 2024
Copy igh : © 2024 by he au ho s.
Licensee MDPI, Basel, Swi ze land.
This a icle is an open access a icle
dis ibu ed unde he e ms and
condi ions o he C ea i e Commons
A ibu ion (CC BY) license (h ps://
c ea i ecommons.o g/licenses/by/
4.0/).
elec onics
Sys ema ic Re iew
Machine Lea ning Fi s Response o COVID-19: A Sys ema ic
Li e a u e Re iew o Clinical Decision Assis ance App oaches
du ing Pandemic Yea s om 2020 o 2022
Goizalde Badiola-Zabala 1,*, Jose Manuel Lopez-Guede 1,2 , Julian Es e ez 1,3 and Manuel G aña 1,3
1Compu a ional In elligence G oup, Basque Coun y Uni e si y (UPV/EHU), 01006 Vi o ia-Gas eiz, Spain;
[email p o ec ed] (J.M.L.-G.); [email p o ec ed] (M.G.)
2
Depa men o Sys ems and Au oma ic Con ol, Facul y o Enginee ing o Vi o ia, Basque Coun y Uni e si y
(UPV/EHU), Nie es Cano 12, 01006 Vi o ia-Gas eiz, Spain
3Depa men o Compu e Science and A i icial In elligence, Facul y o In o ma ics, Basque Coun y
Uni e si y (UPV/EHU), Paseo Manuel de La dizabal 1, 20018 Donos ia-San Sebas ian, Spain
*Co espondence: [email p o ec ed]
Abs ac : Backg ound: The decla a ion o he COVID-19 pandemic igge ed global e o s o con ol
and manage he i us impac . Scien is s and esea che s ha e been s ongly in ol ed in de eloping
e ec i e s a egies ha can help policy make s and heal hca e sys ems bo h o moni o he sp ead and
o mi iga e he impac o he COVID-19 pandemic. Machine Lea ning (ML) and A i icial In elligence
(AI) ha e been applied in se e al on s o he igh . Fo emos is diagnos ic assis ance, encompassing
pa ien iage, p edic ion o ICU admission and mo ali y, iden i ica ion o mo ali y isk ac o s,
and disco e ing ea men d ugs and accines. Objec i e: This sys ema ic e iew aims o iden i y
o iginal esea ch s udies in ol ing ac ual pa ien da a o cons uc ML- and AI-based models o
clinical decision suppo o ea ly esponse du ing he pandemic yea s. Me hods: Following he
PRISMA me hodology, wo la ge academic esea ch publica ion indexing da abases we e sea ched
o in es iga e he use o ML-based echnologies and hei applica ions in heal hca e o comba he
COVID-19 pandemic. Resul s: The li e a u e sea ch e u ned mo e han
1000 pape s;
220 we e
selec ed acco ding o speci ic c i e ia. The selec ed s udies illus a e he use ulness o ML wi h
espec o suppo ing heal hca e p o essionals o (1) iage o pa ien s depending on disease se e i y,
(2) p edic ing admission o hospi al o In ensi e Ca e Uni s (ICUs), (3) sea ch o new o epu posed
ea men s and (4) he iden i ica ion o mo ali y isk ac o s. Conclusion: The ML/AI esea ch
communi y was able o p opose and de elop a wide a ie y o solu ions o p edic ing mo ali y,
hospi aliza ions and ea men ecommenda ions o pa ien s wi h COVID-19 diagnos ic, opening
he doo o u he in eg a ion o ML in clinical p ac ices igh ing his and o ecoming pandemics.
Howe e , he ansla ion o he clinical p ac ice is impeded by he he e ogenei y o bo h he da ase s
and he me hodological and compu a ional app oaches. The li e a u e lacks obus model alida ions
suppo ing his desi ed ansla ion.
Keywo ds: COVID-19; machine lea ning; a i icial in elligence; mo ali y; p edic ion; isk ac o s;
d ug epu posing; d ug
1. In oduc ion
The se e e acu e espi a o y synd ome co ona i us 2 (SARS-CoV-2) was i s epo ed
in Wuhan in Decembe 2019 wi h common clinical symp oms, such as e e , cough, muscle
o body aches, a igue, conges ion o unny nose. The i us was decla ed a po en ial
heal h haza d o people wi h backg ound diseases as i a ec s he uppe -lowe espi a o y
sys em and can cause lung in ec ions and ch onic lung obs uc ion. The Wo ld Heal h
O ganiza ion (WHO) decla ed a pandemic c isis on 11 Ma ch 2020. An ea ly es ima ion
Elec onics 2024,13, 1005. h ps://doi.o g/10.3390/elec onics13061005 h ps://www.mdpi.com/jou nal/elec onics
Elec onics 2024,13, 1005 2 o 38
o he
SARS-CoV-2
a e age ep oduc ion a e (i.e., he a e age numbe o cases o in ec-
ion caused by an iden i ied in ec ed indi idual) was 3.28 [
1
], explaining he pe cei ed
exponen ial g ow h o cases a he e y beginning o he pandemic. The po en ial o
heal h complica ions and he apid sp ead o he i us induced go e nmen s a ound he
wo ld o dic a e s ic popula ion con ol measu es o p e en he sp ead o he i us.
As a esul o he panicked global esponse, demand o inno a i e heal hca e esou ces
inc eased d ama ically. O e he cou se o his pandemic, in ense esea ch has add essed
designing A i icial In elligence (AI), Machine Lea ning (ML) and obo ic solu ions o
imp o e diagnos ic se ices, isk assessmen , moni o ing, and ele-assis ance, aiming o
educe signi ican ly he wo kload o on -line heal hca e wo ke s. Such p essu e on he
scien i ic communi y has gene a ed a eal sunami o publica ions. A sea ch unde he e m
COVID-19 e u ns o e ou -hund ed- housand e e ences in PubMed; his igu e gi es
a good imp ession o he magni ude o he ask o any e iew e o . In his se ing, i is
p oposed ha ML and AI can supplemen da a analysis, o ins ance colla ing candida e
d ug ela ionships, looking o pha maceu ical a ge s in he mids o massi e amoun s
o da a, inding bioma ke s and simula ing he eac ion o main compounds, g ea ly ac-
cele a ing he speed o esea ch and imp o ing quali y and e icacy [
2
]. Hence, ML and
AI p omised g ea po en ial wi h espec o da a-d i en solu ions o help humankind deal
wi h COVID-19 [3].
E en a e he decla a ion o he end o he pandemic by he WHO on 6 May 2023,
COVID-19
is no a ully unde s ood disease, wi h mul iple biological and clinical mani-
es a ions. New symp oms o he disease con inue o be epo ed [
4
], and he guidelines
conce ning which pa ien s should be conside ed Pe sons Unde In es iga ion (PUIs) o
COVID-19 o es ed o SARS-CoV-2 in ec ion ha e been con inually e ol ing du ing he
pandemic yea s. Indeed, clinical needs a e e ised equen ly [
5
]. One o he salien ea u es
o COVID-19 ha has un olded a e se e al yea s in o he pandemic is he e olu ion o he
symp oms o he disease [
6
], o en a ibu ed o he e olu ion o he unde lying i us [
7
],
accompanied by he changes and unce ain ies associa ed wi h he de ec ion and es ing
me hods [
8
]. This unce ain y in he case labeling se iously comp omises he ex apola-
ion o he esul s o Machine Lea ning (ML) algo i hms (i.e., gene aliza ion o ex e nal
alida ion) o di e en pe iods and si es o he pandemic. On he o he hand, some isks
ac o s we e iden i ied ea ly on and emain as such h ough he e olu ion o he pandemic.
Pa amoun isks a e obesi y [9–11] and age [12,13].
P e ious e iews o he use o ML o AI ools in he con ex o COVID-19 a e ei he
e y b oad o deal wi h na ow issues. B oad e iews [
14
–
16
] ackle a wide a ie y o
echniques and issues ha a e some imes un ela ed o clinical decision issues in COVID-19
pa ien s, such as he epidemiological model o disease ansmission [
16
]. O he s deal wi h
he p oblem o disease diagnosis in gene al [
17
]. Na ow e iews deal wi h speci ic aspec s,
such as he e ec s o age [
12
,
13
], obesi y [
10
,
18
], d ug epu posing [
19
], ca dio- ascula
isks [20], analysis o medical images [21,22], and es ing de ices [8].
This s a e o he a has been e iewed ollowing esea ch ques ions ha had been
aised p e iously in o de o c ea e a pa hway o he ealiza ion o a speci ic s udy and
no only o c ea e a e iew o a speci ic opic [
17
,
21
,
22
]. This pape p esen s a sys ema ic
e iew o ML app oaches ha ha e been p oposed in he ea ly pandemic esponse o deal
wi h ele an clinical decision issues: pa ien iage a admi ance, p edic ion o in ensi e
ca e uni (ICU) admi ance, p edic ion o dea h ou come, and iden i ica ion o mo ali y
isks. Addi ionally, we e iew AI-based app oaches o ea men design, speci ically a
sea ch ia AI ools o d ugs a ge ing he COVID-19 pa hogen, SARS-CoV-2.
2. Resea ch Ques ions
The main aim o his pape is o e iew he di e se ML app oaches ha we e p oposed
as a i s esponse o he COVID-19 pandemic in o de o alle ia e he cogni i e and
adminis a i e bu den o heal hca e p o ide s. Mos pa ien s en e a hospi al h ough he
Eme gency Depa men (ED); hence, ML ools may ha e a g ea impac he e wi h espec
Elec onics 2024,13, 1005 3 o 38
o managing he expec ed o e load. Howe e , i is impe a i e o assess he quali y o he
s udies om an ML me hodological poin o iew in o de o asce ain i hese s udies may
be (ha e been) ans e ed e ec i ely and in a imely manne o he clinical p ac ice as
ac ual suppo o he clinician.
This scien i ic objec i e is made conc e e in he ollowing Resea ch Ques ions:
RQ1
Did s udies ollow open science s anda ds? Speci ically, ha e he da a used been
published in open access?
RQ2 Which ML models ha e been mos equen ly p oposed and alida ed?
RQ3
Which a iables/ ea u es a e aken in o accoun and which a e he mos signi i-
can isks ound?
RQ4 Which alida ion p o ocols o ML models ha e been mos equen ly applied?
RQ5
Which pe o mance measu es a e epo ed? Which a e he pe o mances achie ed
acco ding o hese measu es?
3. Me hods
Sea ch S a egy
A sys ema ic li e a u e sea ch was conduc ed in acco dance wi h he P e e ed Re-
po ing I ems o Sys ema ic Re iews and Me a-Analysis (PRISMA) guidelines. Li e a u e
sea ches we e conduc ed in he Web o Science and IEEExplo e si es. The sea ch s ing
con ained he ollowing e ms: “(COVID-19 OR SARS-CoV-2 OR co ona i us) AND (ma-
chine lea ning OR deep lea ning OR a i icial in elligence)”. The JabRe e e ence da abase
manage was used o pe o m he li e a u e sea ch om Janua y 2020 o June 2022. The se-
lec ion p ocess is as ollows: i les and abs ac s o he pape s e ie ed ia he sea ch a e
sc eened, emo ing hose ha do no mee he inclusion c i e ia (no duplica ed, epo ing
ac ual p edic i e pe o mance esul s o e a da ase collec ed in a clinical se ing, no
being epidemiological s udies based on agg ega ed da a, gi ing a de ailed speci ica ion o
eco ded a iables, gi ing de ailed desc ip ions o ML and s a is ical me hods). Selec ed
pape s we e analyzed acco ding o he esea ch ques ions enume a ed abo e.
4. Resul s
4.1. Sea ch Resul s
The bibliog aphic sea ch yielded 2119 esea ch pape s. A e sc eening i les and
abs ac s, 1759 did no mee he inclusion c i e ia. A u he 140 s udies we e excluded in he
nex s age when he ull ex s o his se o a icles we e assessed, lea ing 220 pape s eligible
o analysis. Figu e 1illus a es he pipeline o he pape -selec ion p ocess.
Tables 1–3
summa ize he salien de ails o he selec ed pape s dis ibu ed in o he ollowing clinical
heal hca e opics whe e ML has been p oposed o decision assis ance: (1) iage o pa ien s
a admission, (2) isk o COVID-19- ela ed ICU admission and (3) isk ac o s o COVID-19-
ela ed dea h.
Elec onics 2024,13, 1005 4 o 38
Figu e 1. Flowcha o he pape -selec ion p ocess.
Elec onics 2024,13, 1005 5 o 38
Table 1. S udies epo ing a pa ien iage sys em. Da a segmen a ion (DS), Da a Augmen a ion (DA), Da ase Sou ce access (DSS), T aining Da ase (T ), Tes ing
Da ase (Te), Accu acy (AC), F1-sco e (F1), Sensi i i y (Se), Speci ici y (Sp), Recall (R), C oss-Valida ion (CV), Ex e nal Valida ion (EV), In e nal Valida ion (IV).
Re . Region Model Fea u es DS DA DSS T Te AC F1 Se AUC Sp R Valida ion
[23]Wuhan,
China RF
age, hype ension, ca dio ascula
disease, gende and diabe es o he
clinical ea u es modali y
and dime ized plasmin agmen D,
N/A N/A P i a e 290 72 0.97 0.97 0.99 0.97 0.94 N/A N/A
[24] Beijing Lasso eg ession
Age, Tempe a u e, HR, e e
classi ica ion; headache,
in e leukin-6; sys olic blood p essu e;
monocy e a io; pla ele coun ;
dias olic blood p essu e
N/A Random
o e sampling P i a e 105 27 N/A 0.57 N/A 0.84 0.727 1.00 10- old CV
[25] N/A XGBoos Top 60 impo an ea u es consis ing
o 19 p o eins, 11 me aboli es,
7 lipids, and 23 mRNAs N/A N/A P i a e 108 27 N/A N/A N/A 0.93 N/A N/A 5- old CV
[26] China RF and SVM
28 ea u es (age, gende , whi e blood
cell, neu ophil pe cen age,
lymphocy e pe cen age, monocy e
pe cen age, ...)
N/A Boo s ap
esampling N/A 40 11 0.9 N/A 0.88 N/A 0.9 N/A 10- old CV
[27] N/A BN, NB, MLP,
LWL and RF
age and gende , blood o issue
sample esul s, he pe iod o he
illness, symp oms and lab esul s,
and isk ac o s
N/A N/A N/A 880 587 0.99
(MLP) N/A N/A N/A N/A 0.99
(MLP) N/A
[28]Salamanca,
Spain RF, xgboos
and LR
demog aphic a iables,
como bidi ies, clinical cha ac e is ics,
physical examina ion pa ame e s
and biochemical pa ame e s
a ailable a hospi al admission
N/A N/A N/A 734 184 N/A N/A 0.9 0.83 0.52 N/A
10-s a i ied
old CV
scheme wi h
10 epe i ions
[29] China DL
clinicians o es ima e an indi idual
COVID-19 pa ien isk and make
decisions based on a ailabili y o
esou ces o c i ical pa ien s and
pa ien o e load
N/A N/A N/A 752 188 N/A N/A 0.95 0.894 0.95 0.44 N/A
[30] Ox o d LR, RF and
XGBoos
p esen a ion blood es s, blood gas
es ing, i al signs, and esul s o
PCR es ing o espi a o y i uses N/A N/A N/A 303 77 N/A N/A 0.774 0.939 0.948 N/A 10- old CV
[31]Mad id,
Spain x-means
clus e ing
10 Vi al signs, 29 labo a o y es s and
168 ICD-10 codes N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A
[32] China
RF, GB, SVM, NB,
KNN, LR
Demog aphic da a, como bidi ies,
ou pa ien medica ions, i al signs
and labo a o y alues N/A N/A N/A 80% 20% 0.9 N/A N/A N/A N/A N/A N/A
[33] N/A XGBoos 24 ea u es (a e PCA) N/A N/A P i a e 102 25 0.97 0.96 N/A 1 N/A 0.95 EV
[33,34] B azil N/A ea u es o ou ine blood analysis N/A smo e N/A N/A N/A 0.91 0.87 0.83 0.74 0.91 0.83 EV

Elec onics 2024,13, 1005 6 o 38
Table 1. Con .
Re . Region Model Fea u es DS DA DSS T Te AC F1 Se AUC Sp R Valida ion
[35] Pa ia ResNe 50
age, SBP, DBP, RT, SPO2, empe a u e,
hemogoblin, whi e blood cell,
lymphocy es, Pla ele s, C- eac i e
p o ein and Lac a e dehyd ogenase
N/A
Image noise,
Colou
ji e ing, lip,
cen e c opping
N/A 337 45 0.9872 0.9922 N/A 0.9997 N/A 98.62 15% da ase
[36] N/A Ada Boos , RF,
XGBoos , Ca Boos
age, sex, espi a o y pa ame e s
(SPO2, RR), ca dio ascula
pa ame e s, body empe a u e,
symp oms, associa ed como bidi ies,
ull blood coun ,
biochemical pa ame e s
N/A
b igh ness
changes,
con as
adjus men
and pa allel
shi ing
P i a e 380 95 0.9859 N/A 0.9793 N/A 0.9897 N/A 5- old CV
[37] Mexico NN, LR, SVM,
KNN
APACHE II sco e, whi e blood cell
coun , ime om symp oms o
admission, SPO2 and blood
lymphocy es coun
N/A N/A Public 301,421 64,590 0.81–
0.931 N/A 0.83–
0.961 N/A 0.8–0.92 N/A 15% o da ase
[38]
Cheikh
Zaid
Hospi al,
Mo occo
X_GBoos ,
AdaBoos , RF
and Ex aT ees
Sex. Age. Pla ele , Lymphocy e, PLR,
ALT, AST, LDH, D-dime s,
C_ eac i e p o ein, Weigh ,
Como bidi ies
N/A N/A P i a e 225 97 1 N/A 1 1 1 N/A N/A
[39]JinYinTan
Hospi al,
China RF, SVM, LR
ches compu ed omog aphy, e e ,
malignan umo , HR, SBP,
hemoglobin concen a ion,
neu ophil- o-lymphocy e a io
N/A N/A P i a e N/A N/A 0.845–
0.885 N/A 0.923–
0.967 0.928–
0.970 0.695–
0.79 N/A 10- old CV
[40] N/A RF, XGBoos ,
KNN, MLP, LR,
J48, NB
Gende , age, leng h o
hospi aliza ion, Smoking, ICU
admission, hype ension, pneumonia,
diabe es, ca diac disease, symp oms,
BUN, WBC, C- eac i e p o ein,
hype sensi i e oponin, glucose,
e y h ocy e sedimen a ion a e,
c ea inine, alkaline phospha ase
N/A SMOTE N/A 1350 150 0.9503 N/A 0.907 0.9902 0.951 N/A 10- old CV
[41] N/A DL X- ays, adiology epo s and
RT-PCR da a N/A N/A p i a e 11,599 800 0.77 N/A 0.683 0.925 0.966 N/A
in e nally
(B own-Ap il)
and ex e nally
(Ex e nal and
Xiangya-
Feb ua y)
Elec onics 2024,13, 1005 7 o 38
Table 1. Con .
Re . Region Model Fea u es DS DA DSS T Te AC F1 Se AUC Sp R Valida ion
[42] Tu key
LR, SVM, Vo ed
Pe cep on, KNN,
K s a , LWL, NB,
SGD, DT,
Hoe ding DT, RF
age, da a, lymphocy es coun (LYM),
neu ophils coun (NEU), whi e
blood cells (WBC), mean co puscula
olume (MCV), mean pla ele
olume (MPV) and e y h ocy e
dis ibu ion wid h (RDW),
eosinophils coun (EOS), monocy es
coun (MONO), ed blood cells coun
(RBC), hema oc i , hemoglobin and
(mean co puscola hemoglobin con-
cen a ion (MCVC)
N/A N/A p i a e 3362 840 0.8762–
0.9786 0.9271–
0.988 0.9107–
0.9920 0.8810–
0.9786 0.8762–
0.9786 N/A 10- old CV
[43] N/A E icien Ne CXR images N/A N/A public/
p i a e 455 150 0.8667 0.7865–
0.9174 N/A 0.95 N/A 0.7–1 IV
[44] China LR Age, Sex, Como bidi y, p ima y
symp ons, ou comes,
labo a o y indica o s
lung
images seg-
men a ion N/A N/A 628 158 N/A N/A 0.833 0.732 0.781 N/A 5- old CV
[45] China LR
CK-MB, neu ophils, PCT, α-HBDH,
D-dime , LDH, glucose, PT, APTT,
RDW (SD and CV), ib inogen
and AST
N/A N/A p i a e N/A N/A N/A N/A N/A 0.83 N/A N/A EV
[46]Sou h
Ko ea XGB body empe a u e, pulse a e, RR,
blood p essu e, any symp oms, and
pas medical his o y N/A N/A p i a e 119,576 29,895 0.923 0.861 N/A 0.95 0.933 0.807 N/A
[47] UK RF
Age, Gende , BMI, Smoking S a us,
SPO2, Tempe a u e, como bidi ies,
Albumin, Whi e Blood Coun , Blood
U ea Ni ogen, Lymphocy e Coun
N/A N/A p i a e 1196 299 0.76 0.67 0.78 0.83 0.75 N/A 5- old CV
[48] Ko ea LGBM, ORL
AGE and SEX and BMI, HR,
empe a u e, SBP, DBP, ch onic
ca diac disease, as hma, ch onic
obs uc i e pulmona y disease,
hemoglobin, pla ele s, WBC
N/A N/A p i a e 3,940 1688 0.85–0.88 0.49–0.57 N/A N/A N/A 0.44–
0.56 10- old CV
[49] Is ael Mul is a e Cox
eg ession Age, sex, pa ien being in
1 o 3 clinical s a es N/A N/A N/A 297 33 N/A N/A N/A 0.88 N/A N/A 8- old CV
epea ed
8 imes
[50]
Philippines
DT Sex, Age, Region, N/A Random un-
de sampling N/A 197,164 N/A 0.8142 0.1674 0.8165 0.876 0.8141 N/A 5- old CV
Elec onics 2024,13, 1005 8 o 38
Table 2. S udies p edic ing he ans e o ICU and Leng h o S ay (LoS) in o COVID-19 pa ien s. Da a segmen a ion (DS), Da a Augmen a ion (DA), Da ase Sou ce
access (DSS), T aining Da ase (T ), Tes ing Da ase (Te), Accu acy (AC), F1-sco e (F1), Sensi i i y (Se), Speci ici y (Sp), Recall (R), C oss-Valida ion (CV), Ex e nal
Valida ion (EV), In e nal Valida ion (IV).
Re . Region Model Fea u es DS DA DSS T Te AC F1 Se AUC Sp R Valida ion
[51] Ge many Explainable
Boos ing
Machine 49 a iables N/A N/A N/A 949 237 0.73 N/A N/A 0.69 N/A N/A 5- old CV
[52]Galicia,
Spain
MLP.
DeepNe wo k,
RF, AB,SVM,
KNN, LR
Age, gende , diabe es, hype ophy,
hyd ocele, pneumonia, equen
u ina ion, he apeu ic ad ice, whi e
blood cells, hea ailu e
N/A Smo e, adasyn N/A 110,454 N/A N/A N/A N/A 0.761 N/A N/A 10- old CV
[53] B azil RF, XGB, LR 67 a ibu es N/A N/A N/A 5644 N/A 0.94 0.91 N/A 0.9 0.95 0.92 10- old CV
[54]Wuhan,
China N/A 194 a iables N/A N/A N/A 586 147 N/A N/A >0.571 >0.622 >0.353 N/A 3- old CV
[55] Sou h Asia RF, KNN, SVM Alcoholic be e ages, animal
p oduc s, ce eals excluding bee ,
mea , ege al p oduc s N/A N/A N/A N/A N/A >0.77 N/A N/A N/A N/A N/A N/A
[56]Dubai,
UAE DT Age, gende , na ionali y,
blood g oup, BMI N/A N/A N/A 1513 504 0.96 N/A 0.965 N/A 0.878 N/A 10- old CV
[57] USA XGBoos Age, gende , acu e diagnoses N/A N/A N/A 2313 N/A N/A N/A 0.9 N/A 0.58 N/A N/A
[58]Wuhan,
China XGBoos
lymphocy e pe cen age, p o h ombin
ime, lac a e dehyd ogenase, o al
bili ubin, eosinophil pe cen age N/A N/A N/A 98 25 N/A N/A 0.8 0.92 0.9 N/A 5- old CV
[59] NY, USA RF 31 a iables N/A N/A N/A 1375 612 N/A 0.762 0.728 0.79 0.763 N/A 10- old CV
[60] Denma k RF Age, sex, BMI, como bidi ies,
smoking, lab es s and empo al
ea u es N/A N/A N/A 42,526 N/A N/A N/A N/A 0.995 N/A N/A N/A
[61] NY, USA LR, DT, RF,
GBDT
RT-PCR esul s, ou ine labo a o y
es ing esul s and pa ien
demog aphic in o ma ion N/A N/A N/A N/A N/A N/A N/A 0.761 0.854 0.808 N/A 5- old CV
[62] I an LR, NN, C5.0, RF,
XGBoos
demog aphic cha ac e is ics,
pa ien ’s backg ound, disease
symp oms and a a ge a iable N/A N/A N/A 318 80 0.7901–
0.852 N/A 0.9091–
0.9273 N/A 0.5385–
0.7308 N/A 10- old CV
[63]
Philadelphia,
PA
and P o i-
dence, RI,
USA
RF
demog aphic, clinical and labo a o y
a iables aken on admission o he
ICU including age, sex, empe a u e,
SpO2, WBC, absolu e lymphocy e
coun , se um c ea inine
concen a ion, CRP
and como bidi ies
ches X- ays
we e
segmen ed N/A P i a e 546 108 0.727 0.707 0.714 0.732 0.746 N/A 5- old CV
Elec onics 2024,13, 1005 9 o 38
Table 2. Con .
Re . Region Model Fea u es DS DA DSS T Te AC F1 Se AUC Sp R Valida ion
[64]Kingdom
o Saudi
A abia DL
The da ase con ains demog aphic
ea u es, labo a o y esul s om
comple e blood coun CBC
and adiological indings
and como bidi y
Ches X- ay
segmen a-
ion SMOTE P i a e 1210 303 0.904 N/A 0.86 0.875 0.84 N/A 10- old CV
[65] N/A XGBoos Sex, Ma i al s a us, Age, numbe o
admissions, ype o admissions,
hospi aliza ion da a, como bidi ies N/A N/A N/A 5212 579 0.917 0.918 0.916 0.91 0.913 N/A 10- old CV
[66] N/A
Gaussian mix u e
mode b ea hing equency (BF) and (SpO2) N/A N/A N/A N/A N/A 0.878 N/A N/A 0.94 N/A N/A N/A
[67]
Tel-A i
Sou asky
Medical
Cen e ,
Is ael
Ca Boos demog aphics, backg ound disease,
i al signs and lab measu emen s N/A N/A P i a e 20,029 5007 0.8 N/A N/A 0.76 N/A 0.7
20- old CV and
EV on da a
om a
di e en
hospi al
[68] USA DL
condi ion, p ocedu e, measu emen ,
obse a ion, d ug, de ices, paye ,
isi , heal h p o ide s, medical si es
and pe sonal in o ma ion
N/A N/A Public 9545 200 N/A N/A N/A 0.9 N/A N/A
10 imes by
andomly
spli ing
aining and
alida ion da a
[69]Saudi
A abia J48
demog aphics, como bidi ies, signs
and symp oms o COVID-19 illness,
labo a o y alues,
mechanical en ila ion
N/A N/A P i a e 1101 367 0.731 N/A N/A 0.7542 N/A N/A 10- old CV
[70]Saudi
A abia RF
Age, gende , Weigh , ca ee , Hea
disease, Hype ension, Diabe es
melli us, S oke, Vascula disease,
lymphocy e absolu e alue
N/A N/A p i a e 115 - 0.9416 0.9414 N/A N/A N/A 0.9416 29 samples o
alida ion
[71]Tokyo,
Japan LR Age, gende , como bidi ies, smoing
s a us, empe a u e, RR, SBP, DBP,
HR, SpO2 N/A N/A P i a e N/A N/A 0.9 0.695 N/A 0.875 0.976 0.513 No alida ion
[65] N/A XGBoos Sex, Ma i al s a us, Age, numbe o
admissions, ype o admissions,
hospi aliza ion da a, como bidi ies N/A N/A N/A 5212 579 0.917 0.918 0.916 0.91 0.913 N/A 10- old CV
[72]London,
UK LR, RF and
XGBoos 64 clinical ea u es N/A N/A p i a e 299 145 N/A 0.42–0.60 N/A 0.76–0.87 N/A 0.48 3- old CV
[73]Basque
Coun y,
Spain Ca Boos
p essu e o oxygen, pla ele s,
lymphocy es, monocy es and
eosinophils and he highes alues
o CRP, age, p ocalci onin,
u ea, LDH
N/A N/A P i a e N/A N/A N/A N/A 0.99 N/A 0.95 N/A EV
[74]Lomba dy;
I aly LR age, gende , home medica ions,
como bidi ies and en ila ion
pa ame e s om he i s 24 h in ICU N/A SMOTE P i a e N/A N/A 0.7–0.76 0.71–0.75 N/A 0.77–0.8 N/A 0.72–
0.75 10- old CV
Elec onics 2024,13, 1005 16 o 38
applied andom o e sampling (3%), and 1% o pape s applied o he echniques such as
boo s ap esampling. The emaining 83% o s udies did epo balanced da ase s.
4.3. RQ 2: Machine Lea ning Algo i hms
A majo i y o he selec ed s udies ha e benchma ked models in o de o achie e he
highes accu acy and bes pe o mance acco ding o he p oposed casuis y. The mos
equen ly used algo i hms we e Random Fo es (RF), Logis ic Reg ession (LR), G adien
Boos e (XGB) and Suppo Vec o Machines (SVM) wi h 18.88%, 13.29%, 9.84% and 7.71%
equency, espec i ely. These models a e also he ones ha ha e achie ed he highes
o e all p ecision in compa ison o he emaining ones. The equency o appea ance o ML
algo i hms o e he selec ed pape s is ep esen ed in Figu e 3. Many pape s did y se e al
algo i hms, so ha hese numbe s include o e lapping s udies.
Figu e 3. Numbe o selec ed pape s epo ing esul s o speci ic ML algo i hms. Ac onyms:
RF Random Fo es , LR Logis ic eg ession, GBM, LGBM, GBDT, XGB, XGBoos and Ca Boos a e
compe ing implemen a ions o g adien boos ing, Adaboos Adap i e Boos ing, DT and J48 a e
implemen a ions o decision ees in R and Weka, espec i ely, SVM suppo ec o machines, KNN k-
nea es neighboo , DL and DeepNN a e deep lea ning implemen a ions, MLP mul i-laye pe cep on,
NB nai e Bayes and NN neu al ne wo k.
Among he epo ed models, he LR app oach is he mos deeply oo ed in classical
s a is ics. I is a linea eg ession o a logis ic unc ion ha can be in e p e ed as a pos e io
p obabili y o he posi i e class; he e o e, i allows one o assess he impo ance o each
inpu a iable as an odd a io and he di ec ion o he in luence o he a iables, i.e., i ising
alues inc ease o dec ease he isk. Fo his eason, many s udies applied only LR, also
i he aim o he s udy was o assess he isks o speci ic ac o s like age, gende o o he s.
The emaining app oaches a e ML app oaches, which can p o ide some in o ma ion on
he in luence o impo ance o inpu a iables, bu o en hey canno p o ide he di ec ion
o he o he in luence. The mos popula ML algo i hm is he well known RF which
o en p o ides he bes accu acy esul s and a iable impo ance based on he Gini index
o da a spli s in he ee lea es. Ano he classical app oach is he SVM ha has been
ex ensi ely used in medical s udies o p edic i e model cons uc ion. Howe e , i is no
easy o assess a iable impo ance om SVM, so his aspec is o en omi ed in s udies
exploi ing SVMs. Boos ing app oaches such as Adaboos and di e se la o s o G adien
Boos ing ha e also been popula in COVID-19 s udies abou clinical decision assis ance,
epo ing some imes op imal pe o mances bu lacking he explana o y abili y o a iable
impo ance assessmen . I is impo an o no ice ha di e se ML pla o ms, such as R

Elec onics 2024,13, 1005 17 o 38
packages, Py hon, Weka o Ma lab oolboxes, p o ide sligh ly di e en implemen a ions
o he same models ha may e en appea wi h di e en names. In his e iew, we ha e
no del ed in o his le el o de ail, bu an exhaus i e wo k o compa ison should assess
he di e ences o epo ed esul s o di e se implemen a ions o e he same da ase s.
A i icial Neu al Ne wo ks (ANN), such as he popula Mul i Laye Pe cep on (MLP) o e
a wide spec um o po en ial models wi h many po en ial di e ences in hype pa ame e s
and aining algo i hms. In his s udy, we ound ew Deep Lea ning (DL) app oaches
because da ase s a e o en small ela i e o he la ge da ase s equi ed by DL aining.
Ano he incon enience o ANNs is hei educed explainabili y. Va iable impo ance in
ANNs is a he di icul o es ima e. In his e iew, i is no possible o conclude ha any o
he ML app oaches ound in he e iewed pape s is supe io and should be ecommended
o e he emaining app oaches.
4.4. RQ 3: Fea u es
In some s udies, he numbe o p edic o s is e y high, while o he s gi e only he
numbe o p edic o s wi hou mo e speci ics. Some s udies, which ha e ga he ed many
a iables, decided o conside ML app oaches o es ima e he impo ance o he a i-
ables and hus educe he dimensionali y by some ea u e-ex ac ion o -selec ion p ocess.
None heless, o he a icles ha speci y he a iables employed, hei majo i y did use
a common co e se o a iables. O he a iables we e usually con ex -dependen . Age,
gende , whi e blood cell coun ing, como bidi ies and blood oxygen sa u a ion a e he mos
equen ly conside ed p edic i e a iables. The mos commonly used ea u es and hei
espec i e equencies o appea ance in he s udies a e p esen ed in Figu e 4.
Figu e 4. Dis ibu ion o disc iminan ea u es o e he selec ed pape s.
Elec onics 2024,13, 1005 18 o 38
4.5. RQ 4: In e nal and Ex e nal Valida ion
Da a a e o en spli in o wo o h ee da ase s ( ain, alida ion and es ) when i comes
o alida ion o an ML p edic i e model. In mos o he su eyed s udies, 70–90% o he
da a we e used o ain he models including hype pa ame e selec ion, while he es
se ed o ex e nal es ing. Un o una ely, 16% o he s udies did no gi e any in o ma ion
abou he alida ion p ocedu e ollowed. Al hough using a hi d pa i ion o he da ase
o in e nal alida ion and model hype pa ame e selec ion was p edominan , ex e nal
alida ion was explici ly men ioned only in en s udies. The mos commonly used me hod
o in e nal alida ion was k- old c oss alida ion. Rega ding he numbe o olds, o e 60%
o he s udies epo ed k = 10, almos 26% o he o he s epo ed k = 5, se ing k = 20, k = 8
and k = 3 was epo ed by 2% each and he emaining 9% o s udies epo ed he use o
c oss alida ion bu did no speci y he numbe o olds used.
4.6. RQ 5: E alua ion Me ics
The me ics used o e alua e he pe o mance o he model we e collec ed and a e
compa ed in Tables 1–3. Accu acy, F1-sco e, sensi i i y, AUC, speci ici y and ecall we e he
me ics epo ed in he s udies. Mos s udies epo ed accu acy sco e and AUC, while he
leas used me ic was ecall. The e is s ong a iabili y in he epo ed esul s. Fo ins ance,
he minimum and maximum AUC alues epo ed o models p edic ing mo ali y was
0.997 and 0.57, espec i ely, while accu acy alues anged in he in e al [0.56, 0.99].
5. Re iew o Su eyed App oaches
In his sec ion, we p o ide a discou se e iew o he main app oaches dealing wi h he
iden i ied c i ical clinical issues. Tables 1–3p o ide summa y de ails o he main cha ac e is-
ics o he pape s o iage a hospi al admission, admission o ICU and
dea h, espec i ely.
5.1. Pa ien T iage Me hods
The iden i ica ion o ell ale COVID-19 symp oms [
132
–
134
] and he ea ly wa ning
o COVID-19-posi i e cases [
61
,
135
–
137
] we e he i s s eps owa ds he managemen o
he disease [
138
–
140
]; classi ying pa ien s acco ding o hei se e i y is he nex c i ical
s ep. Nume ous s udies ha e p esen ed models using ML echniques o p edic pa ien
ou comes [
141
] and se e i y assessmen [
142
,
143
] in SARS-CoV-2-in ec ed pa ien s in
di e en egions o he wo ld [
46
,
47
], p e en ing se e e disease p og ession while aiming o
minimize cos s o he pa ien , he heal hca e sys em and socie y a la ge. I was shown ha
i is possible o disc imina e be ween h ee s a es o he pa ien s’ disease e olu ion and ha
i is possible o make accu a e p edic ions o a pa ien ’s hospi al wo kload based on s a is ics
conce ning age, gende and daily clinical s a us (c i ical, se e e o mode a e) [32,49].
Thus, one o he p ima y esea ch goals is guiding hospi al s a wi h alida ed e -
idence gi ing ad ice on he op imal assignmen o limi ed esou ces while imp o ing
pa ien ou comes [28,50]. The e o e, ML and AI ools ha e been p oposed o he c ea ion
o iage assis an sys ems helping o speed up he decision o he admi ance o pa ien s
in o COVID-19 es ic ed a eas and o de e mine which pa ien s will equi e ei he s anda d
o in ensi e ca e [
25
,
37
,
144
,
145
]. In his spi i , online calcula ion ools o ea ly pa ien
iage we e p oposed [
29
]. Some applica ions o an ML-based classi ie aided in excluding
pa ien s ega ding hei se e i y wi hin 1 h o hospi al admission [
30
] on he basis o ou ine
in o ma ion. Some pape s [
38
] epo ed he ex ensi e compa ison o se e al ML models,
such as XGBoos , AdaBoos , RF and Ex aT ees. Some au ho s educed he iage p oblem
o a classi ica ion o pa ien s in o c i ically ill and non-c i ically ill in o de o p io i ize
hose in immedia e need o u gen ca e [
146
], some p oposing RF models [
23
] be ed wi h
mul imodal da a o a a ie y o algo i hms be es ed ( ee-based, unc ion algo i hms and
lazy lea ning algo i hms) o his ask [42].
Classi ica ion o COVID-19 se e i y in o iage ca ego ies has was also ca ied ou
on clinical da a and labo a o y es s ob ained du ing pa ien examina ion in ED [
31
] and
26 blood ou ine indica o s and se e al demog aphic ea u es [
26
] using a speci ically
Elec onics 2024,13, 1005 19 o 38
de eloped RF-SMA-SVM model. A no el SV-LAR model was exhibi ed o iage based on
blood sample ou ine da a [
34
], in line wi h o he app oaches exploi ing blood sample es
da a o ea ly iage [48]
Some o he s udies also ocused on he di e en ial diagnosis o COVID-19 om o he
simila diseases o o de ec pa ien s wi h high isk o u u e lung diseases, such as a
diagnos ic model o aid in he ea ly iden i ica ion o suspec ed COVID-19 pneumonia
pa ien s [
24
] on admission in e e clinics. O he s classi ied he diagnosis o he pa ien s
in o h ee ca ego ies: COVID-19 pneumonia, non-COVID-19 pneumonia and he heal hy
ones [
43
]. S ill, ano he app oach [
35
] p oposed a DL-based sys em o he classi ica ion o
he se e i y o pneumonia conside ing wo se e i y scales. Addi ional wo ks disc imina ed
be ween in luenza H1N1 and COVID-19 pa ien s [27] using an MLP algo i hm.
Many au ho s wo k wi h models p edic ing se e e COVID-19 based on p o eomics
da a [
39
,
40
]; in his line o wo k, C- eac i e p o ein [
147
,
148
], LDH [
149
–
151
],
pla ele s [152,153]
and D-dime [
45
,
154
,
155
] we e ound o be mos associa ed wi h p edic ing he se e i y
o COVID-19.
Ano he line o esea ch ied o de ec di e en le els o se e i y isk o COVID-19
pa ien s based on X- ay imaging [
33
,
36
,
156
]. Some au ho s concluded ha using adiologi-
cal ea u es in conjunc ion wi h blood es s, ea ly iden i ica ion o pa ien s wi h COVID-19,
who a e a isk o disease p og ession, can be achie ed on admission o hospi al [
44
]. Fu -
he mo e, a p ognos ic model using imaging da a [
41
] no only p edic s he se e i y o a
pa ien ’s illness, bu also he ime un il he pa ien encoun e s his o he i s c i ical e en .
5.2. P edic ion o ICU Admission, P og ession and Leng h o S ay
The co ona i us disease ook i s oll on heal hca e sys ems a ound he wo ld, wi h some
pa ien s equi ing leng hy gene al and in ensi e ca e. Unde he p essu e o an unp ece-
den ed bu den on heal hca e sys ems, he e is a need o ools helping decision make s
o plan esou ce alloca ion a he uni , hospi al and na ional le els, which can be ackled
wi h ML me hods [
157
]. In essence, clinicians will be in e es ed in ML app oaches ha
a e able o p edic whe he pa ien s diagnosed wi h COVID-19 will equi e di e en le els
o hospi al ca e ( ans e om he basic hospi al o ICU) [
52
,
53
,
158
]. Fo ins ance, i is
desi able o eliably p io i ize pa ien s wi h COVID-19 who a e a isk o needing ans e
o he ICU wi hin he nex 24 h, which has b oad implica ions and u ili y o clinical p ac ice
and hospi al ope a ions, on he basis o isks o espi a o y ailu e, shock, in lamma ion
and enal ailu e in he p og ession o COVID-19 [59]. Dealing wi h child en [159] is e en
mo e di icul , because he e y low p e alence o se e e COVID-19 means e y sca ce
da a. Despi e hese d awbacks, some esea che s demons a ed a DL app oach using a
la ge ea u e se o p edic bo h he isk o hospi aliza ion o in ec ed child en and he isk
o se ious complica ions in hospi alized pedia ic pa ien s [
68
]. Ano he line o esea ch
deals wi h he p edic ion o ICU wo kload o e a sho - e m ime ho izon [60] in o de o
a ain op imal esou ce managemen .
Ex ensi e e alua ions o ML algo i hms belonging o nine ca ego ies ha e been e-
po ed o p edic ion o clinical ou comes such as ICU admission o mo ali y, using da a
a ailable om he ini ial COVID-19 de ec ion [
160
]. Speci ically, ICU admission has been
p edic ed on he basis o demog aphic da a [
161
] and como bidi ies [
162
] and some ea ly
symp oms [
163
]. Accu a e assessmen o ICU admission, ICU leng h o s ay and mo ali y
o COVID-19 pa ien s o op imal alloca ion o ICU esou ces has also been epo ed [
54
].
The Explainable Boos ing Machine analysis [
51
] enables one o pe o m p edic i e modeling
o COVID-19 in he ICU as well as he iden i ica ion o isk ac o s. La ge and exhaus i e
clinical da abases ex ac ed om he Elec onic Heal h Reco ds (EHR) allow he applica ion
o DL app oaches o p edic he p obabili y o ICU admission and mo ali y [
164
]. In a
qui e di e en se ing, beha io al AI echniques ha e been also applied o assis in ensi is s
in dealing wi h he decisions conce ning ICU eligibili y [77] in Denma k.
ML me hods can be used o gain knowledge abou complex clinical si ua ions an icipa -
ing u u e complica ions [
74
]. In his ega d, op imally p edic ing mo ali y o COVID-19
Elec onics 2024,13, 1005 20 o 38
pa ien s in he ICU has been epo ed wi h DL echniques [
63
] and ML [
58
], while decision
suppo ools applicable o c i ically ill pa ien s wi h COVID-19 a high isk o 28-day
mo ali y in he ICU allowed assis ance in c i ical decisions such as end-o -li e decisions
and bed alloca ion in cases o limi ed ICU capaci y [
69
]. Fu he mo e, supe ised bina y
p edic ion classi ica ion using a ime-sliding window-based app oach o p edic he isk o
in uba ion 72 h om he end o he 24 h sampling pe iod has been epo ed [
165
]. Mul iple
s udies sugges ed ha adop ing ea ly measu es o ea pa ien s a isk o de e io a ion
could p e en o dec ease s a us wo sening and he need o mechanical en ila ion; in his
way, e e ence [
67
] epo s a model ha p edic s he isk o de e io a ion o each hou ,
while he goal o he s udy in [
73
] was o moni o pa ien p og ession o a sco e o 5 o
mo e on he WHO Clinical P og ession Scale be o e hey equi e mechanical en ila ion.
COVID-19 hospi al eadmission has been ano he subjec o esea ch in e es , due o
i s impac on he op imal managemen o hospi al se ices. ML models can success ully
p edic COVID-19 eadmission [
65
,
75
,
166
]. Conside ing isk ac o s, hese wo ks also
ca ego ized cases wi h a high isk o ein ec ion in o de o classi y pa ien s, making he
u iliza ion o hospi al esou ces mo e e icien . Al e na i ely, o he wo ks [
55
] ocused he
eco e y o COVID-19 pa ien s on die a y adap a ion by pe o ming an analysis o he
ene gy in ake o di e en ood ca ego ies om di e en coun ies, compa ing se e al ML
algo i hms o p edic he eco e y a e.
Al e na i ely, he moni o ing o pa ien s in need o espi a o y suppo would p o ide
goal-o ien ed ools o pa ien isk s a i ica ion and an ale sys em o sel -ca e pa ien s, de-
ec ing highly dis essing s a es when esou ces a e possibly cons ained.
Bu dick e al. [57]
ocused on an ML algo i hm o help e icien iage o pa ien s and esou ce alloca ion by
assessing en ila ion needs among COVID-19 pa ien s, achie ing a p ecise p edic ion o
he mechanical en ila ion esou ce needs wi hin 24 h. Bolou ani e al. [
167
] p oposed an
ML model ha p edic s espi a o y ailu e wi hin 48 h o ED admission. Izadi e al. [
168
]
aimed o iden i y pa ien s who may be a inc eased isk o se e e COVID-19 ou comes due
o he onse o acu e espi a o y dis ess synd ome.
Based p ima ily on pa ien age and measu es o oxygena ion s a us du ing he ED s ay,
i was possible o iden i y pa ien s wi h high isk o poo ou comes, i.e., hose equi ing
in ensi e ca e, hose equi ing mechanical en ila ion and hose wi h high in-hospi al
mo ali y isk [
72
]. The s udy o Saada mand e al. [
62
] p edic ed he equi emen o
oxygen-based ea men o hospi alized COVID-19 pa ien s. Simila ly, Aslam [
64
] iden-
i ied he impac o pa icula a ibu es on he p edic ion o mo ali y and mechanical
en ila ion suppo [
66
] in COVID-19 pa ien s. Iga ashi e al. [
71
] in oduced a model ha
can be implemen ed as a iage ool o de ec he need o supplemen al oxygen.
Finally, unde s anding ha COVID-19 hospi aliza ion imes a e o en long and may
a y subs an ially om pa ien o pa ien , some wo ks [
56
,
70
,
76
] aimed o de elop a eliable
p edic ion model o ED leng h o s ay o COVID-19 pa ien s and o iden i y clinical ac o s,
such as age and como bidi ies, associa ed wi h ED leng h o s ay. AI echniques ha e been
shown also o p edic hospi al occupancy [169] and ICU admission [170].
5.3. Mo ali y P edic ion
In he hea o he pandemic yea s, heal hca e p o essionals o en complained o hei
limi a ions wi h espec o de e mining wi h some p ecision he p ognosis o pa ien s wi h
COVID-19 om he momen o admission h ough o subsequen phases. I has been
epo ed ha he cou se o COVID-19 su e e s unexpec ed changes so ha appa en ly
s able pa ien s suddenly wo sen. In hese si ua ions, e en he mos expe ienced clinicians
may be unable o adjus and espond in a imely manne o he new si ua ion. Thus, ML
and AI models we e p oposed o clinical decision-making, helping de ec complex pa e ns
in la ge da ase s [
20
,
92
,
171
,
172
]. Se e al s udies ha e shown he capabili y o ML-based
models o p edic mo ali y a he le el o indi idual pa ien s [
108
,
115
,
119
,
121
], and he
agg ega e le el o ci ies [
119
]. Rou ine clinical a iables ha e been shown o p o ide
enough p edic i e powe o pa ien
mo ali y [105,173],
while o he s udies ha e shown a
Elec onics 2024,13, 1005 21 o 38
capaci y o p edic he mo ali y o new i us a ian s [
174
] o he e ec o como bidi ies
in COVID-19 mo ali y [120].
Applica ion o AI ools, namely DL a chi ec u es, o medical imaging, speci ically
CT scans, which a e he ones mo e o en done on pa ien s wi h pneumonia complica-
ions, ha e been shown o be use ul o he diagnosis and p ognosis o di e se lung
a ec ions [
175
–
177
]. This app oach also has alue as a p elimina y sc eening ool aiming o
diminish he wo kload on hospi al s a and educe he a e o misdiagnosis o pa ien s wi h
COVID-19 [178–180].
The enhanced p edic ion o disease se e i y ia AI on CT images al-
lows imp o ed mo ali y p edic ion [
126
,
129
,
181
–
187
] and disc imina ion om o he o ms
o pneumonia no due o SARS-CoV-2 [
188
–
190
]. Some wo ks [
79
,
117
] c ea ed a adiomics-
and DL-based model showing he obus ness o he app oach on da a om se e al si es.
Mo eo e , AI allows one o e icien ly combine medical imaging in o ma ion wi h o he
sou ces o clinical and labo a o y in o ma ion [
100
,
106
,
191
]. Fo ins ance, Lu e al. [
192
]
demons a ed o e CT scan cha ac e is ics ob ained ia AI analysis ha he e is a posi i e
co ela ion be ween blood glucose le el on admission and lung lesions.
COVID-19 mo ali y has also been p edic ed on he basis o p o eomics da a. Yasa e al. [
87
]
epo ed an associa ion o a ia ions in blood p o eins wi h he se e i y o he pa ien ’s
condi ion. Addi ionally, Chen e al. [
127
] de eloped an ML p ocedu e o ind bioma ke s
ha de e mine disease se e i y in indi idual immune cells.
O he app oaches use he e ogeneous a iable selec ion om se e al clinical domains.
Azna -Gimeno e al. [
109
] c ea ed an easy- o-use web applica ion ha suppo s apid
decision-making in clinical p ac ice h ough he cons uc ion o a p edic ion model om
a la ge amoun o da a om se e al pandemic wa es ha p edic s ICU admi ance and
mo ali y. Ga a a e al. [
89
] ocused on an ea ly wa ning model based on demog aphic
and clinical a iables o p edic in-hospi al mo ali y o pa ien s wi h COVID-19 in he
ED. Laino e al. [
130
] applied ML echniques o guide he managemen o pa ien s wi h
COVID-19 by de eloping an accu a e in-hospi al mo ali y isk sco e o COVID-19 based
on en a iables. Vezzoli e al. [
124
] p esen ed a isk sco e o in-hospi al mo ali y whe e
mo e se e e pa ien s we e olde , had a lowe blood oxygena ion, lowe c ea inine clea ance
le els and highe p e alence o ele a ed oponin.
Domínguez-Olmedo e al. [
88
] de eloped a model able o p edic mo ali y in pa ien s
wi h COVID-19, which allows one o assess mo ali y om labo a o y alues wi h high
p ecision using he XGBoos model. Sanka ana ayanan e al. [
96
] p esen ed an app oach
using GRU-D ex e nal NN, p o iding an ale sys em o lag mo ali y o COVID-19-
posi i e pa ien s using clinical a iables and labo a o y esul s in a 72 h pe iod a e he
i s posi i e PCR es esul . Halasz e al. [
78
] de eloped an ML-based sco e o 30-day
mo ali y p edic ion in pa ien s wi h COVID-19 pneumonia. Ko e al. [
111
] de eloped an
AI model, EDRne , ha o ecas s he mo ali y a e o COVID-19 pa ien s based on 28 blood
bioma ke s and pa ien age and sex. Vepa e al. [
193
] ob ained da a including age, gende ,
e hnici y, SPO2, RR, empe a u e, obesi y [
18
], as hma, diabe es and hype ension, among
o he s, o a pe iod o almos wo mon hs. The esul s sugges ha low albumin, ele a ed
(CRP) and olde age all co ela e wi h mo ali y in hospi alized pa ien s.
Some esea che s ha e conside ed s udying he geog aphic and demog aphic di e -
ences in luencing bo h sp ead and mo ali y ac oss geog aphical loca ions.
Fidan e al. [84]
employed clus e ing echniques o de e mine ci ies wi h simila isk le els, analyzing he
incidence o cases and en i onmen al pa ame e s. Guzmán-To es e al. [
194
] poin ed
ou ha condi ions in each coun y may di e depending on di e en ac o s such as he
gene al heal h s a us o he people, epo ing ha he main causes o dea h in Mexico a e
ela ed o age, poo ea ing habi s, ch onic diseases and con ac wi h in ec ed people who
do no ha e adequa e ca e. Zawbaa e al. [
82
] p oposed a compa ison o he sp ead o
he disease among nine di e en coun ies, e ealing ha a e age young age, ho clima e,
p e alence o Bacillus Calme e–Gué in (BCG) accine and mala ia ea men a e c ucial
elemen s dec easing he mo ali y impac o he i us.

Elec onics 2024,13, 1005 22 o 38
5.4. Iden i ica ion o Mo ali y Risk Fac o s
Casi aghi e al. [
107
] epo ed a sys em p ima ily designed o ex ac he mos el-
e an adiological, clinical and labo a o y a iables o imp o e pa ien isk p edic ion
and subsequen ly p o ide decision c i e ia o clinicians o suppo pa ien isk assessmen .
Dabbah e al. [
110
] documen ed se e al new and signi ican p edic o s o mo ali y in
COVID-19; hese included de ailed an h opome y, acu e enal ailu e, u ina y ac in ec-
ion and pneumonia. In con as , Baqui e al. [
90
] conside ed demog aphic a iables as he
mos ele an ones, namely he s a e o esidence and i s de elopmen index, he dis ance
o he hospi al (especially o u al and less de eloped a eas), he le el o educa ion and
he inancing model o he hospi al, and s ain. The model p esen ed by Hu e al. [
104
]
could be aluable o clinicians who ha e o iden i y pa ien s a high isk o dea h, based
on age, high-sensi i i y C- eac i e p o ein le el, lymphocy e coun and dime le el, so ha
in e en ions can be aken a an ea lie s age o educe he isk o mo ali y in hese pa ien s.
Fo So oudeh e al. [
83
], he mos impo an ea u es in e ms o p edic ion capaci y,
ou o he many, we e SPO2 (LDH), age, BUN, base excess, c ea inine and WBC. Mah-
da i e al. [
80
] epo ed p omising esul s based on a educed numbe o ea u es such as
SPO2, age and ca dio ascula diso de s u ilizing SVM. The ele ance o LDH is also ound
in he wo k o K ysko e al. [
86
], whe e high le els o LDH, IL-6, IGM, D-dime , ib inogen
and glucose s ongly in luences he se e i y o cases.
Simila o Wang e al. [
93
] in hei inding ha lymphocy e and neu ophil le els in
pe iphe al blood could ep esen p ognos ic ma ke s indica ing ea ly wa ning implica ions,
Li e al. [
99
] epo ed GBDT was e ealed as he mos e ec i e model in e ms o pe o -
mance whe eby leukomonocy es, u ea, age and SPO2 u ned ou o be he mos signi ican
p edic o s. Lip ak e al. [
118
] iden i ied he ele a ed AST as he mos impo an p edic o
o COVID-19- ela ed hospi aliza ions.
Blagoje ic e al. [
81
] ocused on de ec ing he op 10 blood es cha ac e is ics ha a e
s ongly co ela ed wi h pa ien condi ion and based on hese cha ac e is ics p edic ing he
se e i y and clinical s a us. Addi ionally, ecen s udies ha e indica ed ha e i in le el
is conside ed a s ong bioma ke o he p ognosis o SARS-CoV-2 mo ali y [
101
], along
wi h o he se e e espi a o y diseases. Ano he equen ly employed a iable is he le el
o blood sa u a ion [112,113,116].
The s udy o Zhao e al. [
103
] showed ha , combined wi h he pe cen age o neu-
ophils, alanine amino ans e ase, gende and albumin would be associa ed wi h su i al
in pa ien s wi h COVID-19. Age [
122
,
128
] and gende [
123
] appea o be he mos signi -
ican p edic o s o mo ali y [
195
], bu in e ms o symp om–como bidi y combina ions,
Pneumonia–Hype ension, Pneumonia–Diabe es and Acu e Respi a o y Dis ess Synd ome
(ARDS)–Hype ension showed he mos signi ican associa ions wi h
COVID-19
mo al-
i y [
114
]. The s udy by Reina e al. [
125
] examined he hypo hesis ha p e-exis ing condi-
ions (como bidi ies) o pa ien s may inc ease he se e i y o pa ien s due
o SARS-CoV-2.
The C- eac i e p o ein has been consis en ly implica ed in nume ous s udies as a
signi ican indica o o pa ien ou come [
95
,
97
,
98
]. This pa ame e was also aken in o
conside a ion by Boo h e al. [
102
], in conjunc ion wi h blood u ea ni ogen, se um cal-
cium, se um albumin and lac ic acid o he p edic ion o mo ali y in pa ien s up o
48 h p io o dea h. Fang e al. [
94
] c ucially conside ed y oponin, b ain na iu e ic pep-
ide, whi e blood cell coun , aspa a e amino ans e ase and c ea inine, oge he wi h he
abo e-men ioned p o ein as indica o s con ibu ing o malignan p og ession. The com-
bina ion o C- eac i e p o ein and ad anced age, o example, we e he main ac o s o
Di Cas elnuo o e al. [91].
5.5. T ea men s and D ugs
Gi en he high cos in ime and esou ces o d ug disco e y, ML and AI
p oposed
[196–198]
o accele a e he d ug-disco e y p ocess by using in o ma ion abou
he biological, chemical and spa ial p ope ies o compounds and hei po en ial a ge s;
speci ically, COVID-19 DL has been s udied [
199
–
201
]. Thus, he COVID-19 pandemic is
Elec onics 2024,13, 1005 23 o 38
an oppo uni y o bio ech, pha maceu ical and AI companies o coope a e in accele a ing
hei esea ch and de elopmen o iden i y new and a e d ug molecules as well as pe son-
alized medicine [
202
]. Gaw iljuk e al. [
203
] p oposed a p ocess o mul iple i e a ions o ML
models as a p io i iza ion ool o disco e y p og ams o an i i al d ugs agains COVID-19.
P o eomics analysis o plasma om COVID-19 pa ien s e sus con ols led ia SVM ea u e
selec ion o he iden i ica ion o bioma ke s, which can be use ul o d ug a ge ing and
epu posing ia in silico p o ein docking [
204
]. Some wo ks use indi ec in o ma ion abou
he d ug’s e ec , such as ea u es o he images o cells ea ed wi h speci ic d ugs [
205
], so
clus e ing hese images may e eal simila i ies in he mode o ac ion o he d ugs. O he
indi ec in o ma ion used is he cons uc ion o a g aph ep esen a ion o he biological
ne wo k o COVID-19 ha ela es o he a ge s o he i us, making i easy o selec he
essen ial p o eins in he biological ne wo k. In his app oach, e . [
206
] ound i e g oups
o sui able d ugs ha con ain some candida es as po en ial ea men s o COVID-19.
DL app oaches ha e been deployed [
207
,
208
] o ain he molecula desc ip o da ase
o obus d ug disco e y and ea u e ex ac ion o comba COVID-19. Ea ly on in he
pandemic, some esea ch eams ocused on building he in as uc u e o a global e o in
his di ec ion [
209
,
210
]. The applica ion o DL app oaches o e chemical in o ma ion allows
one o ob ain high-le el ep esen a ions o ins ances om a public d ug bank da abase
as well as SARS-CoV-2 amino acid sequences p edic ing a ini y sco es ha ag ee wi h
he SARS-CoV-2 inhibi o s unde e alua ion a he ime o w i ing he pape [
211
]. O he
au ho s p opose o eso o a omic-le el simula ion in o de o ob ain he inpu o DL
app oaches [212].
Repu posing is he use o an exis ing d ug, usually a gene ic and low-cos one, o he
ea men o a new disease o a disease di e en om he one he d ug has been app o ed
o . Repu posing cu en d ugs o he ea men o COVID-19 p o ides a e y apid
pa hway o po en ially e ec i e ea men de elopmen [
19
,
213
–
215
], bo h because new
d ugs a e no equi ed o be de eloped and because, o many exis ing d ugs, sa e y and
e icacy has al eady been es ablished based on p e ious ials. Fo ins ance, some au ho s
claim ha Nai e Bayes achie ed ela i ely high accu acy (72%) p edic ing he e icacy o
epu posing d ugs o COVID-19 [
216
]. The analysis was conduc ed in silico, based on he
published sequences o some p o eins o he SARS-CoV-2 i us. Thus, hey do no ake in o
accoun he e olu ion o he i us and he po en ial o changes in he a ge p o eins. This
is a gene al issue o AI app oaches wi h espec o d ug epu posing o COVID-19.
Some au ho s ha e compa ed p edic ion models o D ug Ta ge In e ac ions (DTIs),
exploi ing he public D ugbank da abase and success ully iden i y in e ac ions be ween
d ugs and p o eins in he human cell [
217
,
218
]. Jin e al. [
219
] de eloped a DL app oach
using ComboNe o he p edic ion o chemical syne gy agains SARS-CoV-2, achie ing
a es ROC-AUC o 0.82, consis ing o an NN a chi ec u e ha join ly lea ns d ug– a ge
in e ac ions and d ug–d ug syne gy.
Bu dick e al. [
220
] ocused on p ecision medicines, as hey may be use ul in iden i ying
a sub-popula ion o COVID-19 pa ien s mos likely o bene i om hyd oxychlo oquine
ea men in a clinical ial. Zeng e al. [
221
] u ilized a ne wo k-based deep-lea ning
me hodology o d ug epu posing, connec ing d ugs, diseases, p o eins and pa hways.
The lack o eliable da a abou he e ec o a d ug on a new disease was esol ed in [
208
]
ia a no el da a-augmen a ion app oach ha exploi s esul s om ailed expe imen s
ha eed a g aph neu al ne wo k. Su e al. [
222
] applied mul iple nonnega i e ma ix
ac o iza ion app oaches o ea u e ex ac ion o i us–d ug associa ion, d ug chemical
s uc u es and i us genome sequences as a p ep ocessing s ep o g aph neu al ne wo k
p ocessing ha ob ains d ug– i us a ini ies. A simila ma ix- ac o iza ion p ocess o
ea u e ex ac ion coupled wi h g aph neu al ep esen a ions was p oposed by [
223
]. O he
esea che s applied a ans e lea ning app oach [
224
,
225
] using a p e- ained DL-based
d ug– a ge in e ac ion model o iden i y comme cially a ailable d ugs ha could ac
on
SARS-CoV-2
i al p o eins. Ke e al. [
226
] p oposed a compa ison o wo models o
de ec d ugs wi h po en ial an i i al e ec s, showing ha adequa e combina ion wi h
Elec onics 2024,13, 1005 24 o 38
o he d ugs a a lowe dose o gemci abine can o e come ad e se pulmona y e ec s while
simul aneously inhibi ing SARS-CoV-2. Yang e al. [
227
] epo ed ha Ca hepsin L (CTSL)
is a p o ease ha can ac i a e he p o ein ha leads o SARS-CoV-2; ne e heless, as he e
is s ill a lack o clinically a ailable CTSL inhibi o s, hey de eloped a DNN app oach o
iden i y small molecules and FDA-app o ed d ugs ha can block CTSL ac i i y o d ug
de elopmen and epu posing o COVID-19.
Mekni e al. [
228
] ocused on SVM me hodology o he p edic ion o inhibi o y ac i i y
o no el chemo- ypes agains SARS-CoV-2, showing an accu acy o 0.88 o la e u iliza ion
o p edic he inhibi o y ac i i y o compounds comme cially a ailable. Kowalewski e al. [
229
]
iden i ied nasal ca i y and espi a o y ac condi ions as a bo leneck o his in ec ion,
ained ML models o p edic inhibi o y ac i i y and sc eened o e 100,000 app o ed d ugs
and mo e han 14 million pu chasable chemicals o po en ial candida es o new inhaled
he apies. Al e na i ely, Pin o e al. [
230
] used mul i a ia e s a is ical me hods o selec he
mos sui able candida es o inhibi he disease.
Finally, u he clinical ials may depend on biomedical in o ma ion ob ained om
exis ing da a and ind pa e ns and signa u es in he unde lying molecula biology o he
COVID-19 mechanism and ML may ha e a p ominen ole in he design o new ials [
231
].
6. Discussion
DQ1. A e p edic i e models capable o suppo ing a COVID-19 ou b eak and how?
ML algo i hms can nei he be adequa ely ained no alida ed wi hou a la ge and
well cu a ed clinical da ase . In he ea ly s ages o an eme ging in ec ious disease, collec ed
da a will o en be noisy and incomple e. Al hough he e may no be a la ge his o ical
da ase o pa ien s su e ing om his disease, exis ing da abases, assuming hey ha e
been p ope ly mined, and pa e ns based on hese da a can con ibu e signi ican ly o he
choice o he mos app op ia e beha iou o adop in each si ua ion [
232
]. Many s udies
epo he po en ial o he applica ion o Big Da a and AI echnology o con ibu e o
p e en ion, diagnosis, ea men and decision making conce ning acu e in ec ious public
heal h e en s in he u u e. Acco ding o some au ho s, he COVID-19 pandemic has
p o ided an excellen oppo uni y o in eg a e AI ools in clinical ca e, al eady in oducing
changes in hospi al p ac ices [233] in some ad anced coun ies.
DQ2.
A e he e demog aphic and cul u al ac o s in luencing he de elopmen o p edic-
i e me hods o con on o add ess COVID-19?
Al hough mos o he a icles added o his s udy deal wi h da a om Asian pa ien s,
he e a e se e al a icles om Eu opean and Ame ican coun ies. In compa ing hese
s udies, while many u ilise simila a iables, cul u al o demog aphic ac o s ha e been e-
ealed o signi ican ly a ec he de ec ion o inc ease in disease con ac ion, demons a ing
he impo ance o hese ac o s in he COVID-19 ou b eak [
173
].In addi ion, key gene ic
ma ke s may se e as po en ial a ge s in he clinical p ognosis and ea men o COVID-19.
DQ3.
A e he e models wi h good pe o mance capable o ca ego izing pa ien s acco ding
o se e i y?
ML has p o en i s impo ance in nea ly all domains and i s echniques a e being
ac i ely used agains COVID-19 by esea che s wi h sa is ac o y esul s [
234
]. Among all
he a icles collec ed in his s udy, a ious s udies wi h e y p omising esul s ha e been
epo ed. Mos o he selec ed s udies employed p edic i e e e enced models o achie e
he highes accu acy and bes pe o mance acco ding o he p oposed casuis y. Among all
epo ed me ics wi h success ul esul s, he e a e se e al s udies ha ha e achie ed
accu acy and AUC abo e 0.99 [
27
,
35
,
60
], showing high sensi i i y alues so ha he mos
se e e cases may no easily go unno iced and can be de ec ed in ime. These a e s udies
ha use MLP, ResNe 50 o RF models and ob ain e y good esul s wi h demog aphic and
clinical a iables, wi hou he need o labo a o y es s o CT images, which a e a iable
and mo e di icul o ob ain and collec .
Elec onics 2024,13, 1005 25 o 38
DQ4.
Is ansla ion possible om scien i ic esea ch o clinical p ac ice wi h he cu en
da a on he disease ob ained du ing he pandemic?
Da ase s used o diagnosis, p edic ion and p e en ion o COVID-19 a e essen ially
classi ied in o medical imaging da ase s, speech-based da ase s and ex ual da ase s. On he
one hand, medical imaging da ase s, i.e., CT images, a e mainly analyzed o au oma ic
diagnosis, segmen a ion and augmen a ion o COVID-19. On he o he hand, ex ual da a
suppo p edic ion and analysis o COVID-19 cases, su eying pa ien s a us, in e en ions
and ea men s.
Based on he s udy, de eloped models add essing COVID-19 h ough in elligen
app oaches gene a e eliable pe o mance esul s i high ideli y and abundan da a a e
in ol ed. I is clea ha la ge da ase s a e no ypically used o he majo i y o he use
cases p e iously speci ied. Despi e he ac ha some public da ase s a e a ailable o
wo k wi h his ype o caseload, gi en he ecen eme gence o he disease and he lack o
many yea s o s udy, he size o hese da ase s is limi ed compa ed o he equi emen s o
ML app oaches.
Da a accessibili y and openness is conside ed a c i ical bo leneck in COVID-19 e-
sea ch as a esul o he apid sp ead. Being able o apply such applica ions in he eal wo ld
will only be possible wi h he a ailabili y o mo e open sou ce da a. This is conside ed one
o he limi a ions o be add essed in o de o u he imp o e COVID-19 esea ch.
This limi a ion p ima ily s ems om he seg ega ion o da a a na ional, egional,
hospi al and depa men al le els. Hence, he de elopmen o a cen al and uni o m
pla o m o in es iga o s o sha ing and accessing da a would be a p omising s a ing
poin . In addi ion o ha ing e oneous o uns uc u ed da a, he da ase s ha e a lo o noise
and null alues. The e o e, il e ing, cleaning and noise educ ion a e o he key challenges
o he success ul implemen a ion o models de eloped using in elligen app oaches.
A u he conside a ion ha mus be add essed is he es ablishmen o ce ain p o ocols
o s anda ds o da a collec ion in heal hca e acili ies. As well as ha ing e oneous o
uns uc u ed da a, he da ase s ha e a lo o noise and missing alues. The e o e, il e ing,
cleaning and noise educ ion a e o he impo an challenges o be conside ed o he
c ea ion o applica ions using in elligen app oaches.
7. Insigh s in o ML/AI Ad ances, Resea ch Di ec ions and Challenges
A he eques o e iewe s, we in oduce he e some hough s abou he u u e o
ML/AI applica ions in clinical decision suppo o COVID-19 and o he u u e pandemics.
Mode n ML and AI echniques a e da a d i en; hus, he main and o emos issue in he
use ulness o ML/AI echniques is he a ailabili y o la ge high-quali y da ase s. I is impos-
sible o o e -s ess he impo ance o good da a-ga he ing p ac ices o he de elopmen o
ML/AI suppo o he esponse o o hcoming pandemics. In con as , he p ice o be paid
by complacency and missing oppo uni ies o implemen a ion o s ong da a-ga he ing
p o ocols will be a g ea e economic dis up ion and poo e heal h ou comes, i.e., mo e
dea hs. Cu en Elec onic Heal h Reco d (EHR) sys ems a e dispe sed and do no in e op-
e a e ou side some small clus e s de eloped by he same company. Hence, in e na ional
s anda ds o EHR should be en o ced as well as da a-ga he ing p o ocols and open access
o da a in o de o allow o coope a i e analysis o da a in o de o aise ala ms and o
p oduce imely p edic i e models and isk assessmen s. Da a access should no be blocked
by comme cial in e es s in he name o da a p i acy when i is mos needed o democ a ic
public heal h managemen . Wi h he excep ion o some p ominen pe sonages, heal h da a
a e easily anonymized o a le el ha impedes spu ious uses.
Rega ding ML/AI models, we hink ha LR will emain one o he mos used models
in he u u e, because i s ounda ions a e well unde s ood and i has high explana o y
alue. Medical p ac i ione s a e well acquain ed wi h he echnique and hey accep i
wi h c i ical app aisal. O he s a is ical ML models, such as RF, SVM o g adien boos ing
app oaches, do no p o ide a simila deg ee o explana ions. They can be o alue in some
speci ic p oblems in o de o p o ide mo e accu a e p edic ions once he isk assessmen
Elec onics 2024,13, 1005 32 o 38
92.
Tezza, F.; Lo enzoni, G.; Azzolina, D.; Ba ba , S.; Leone, L.; G ego i, D. P edic ing in-Hospi al Mo ali y o Pa ien s wi h COVID-19
Using Machine Lea ning Techniques. J. Pe s. Med. 2021,11, 343. [C ossRe ]
93.
Wang, X.; Che, Q.; Ji, X.; Meng, X.; Zhang, L.; Jia, R.; Lyu, H.; Bai, W.; Tan, L.; Gao, Y. Co ela ion be ween lung in ec ion se e i y
and clinical labo a o y indica o s in pa ien s wi h COVID-19: A c oss-sec ional s udy based on machine lea ning. BMC In ec . Dis.
2021,21, 192. [C ossRe ] [PubMed]
94.
Fang, C.; Bai, S.; Chen, Q.; Zhou, Y.; Xia, L.; Qin, L.; Gong, S.; Xie, X.; Zhou, C.; Tu, D.; e al. Deep lea ning o p edic ing
COVID-19 malignan p og ession. Med. Image Anal. 2021,72, 102096. [C ossRe ] [PubMed]
95.
Guan, X.; Zhang, B.; Fu, M.; Li, M.; Yuan, X.; Zhu, Y.; Peng, J.; Guo, H.; Lu, Y. Clinical and in lamma o y ea u es based machine
lea ning model o a al isk p edic ion o hospi alized COVID-19 pa ien s: Resul s om a e ospec i e coho s udy. Ann. Med.
2021,53, 257–266. [C ossRe ]
96.
Sanka ana ayanan, S.; Balan, J.; Walsh, J.; Wu, Y.; Minnich, S.; Piazza, A.; Osbo ne, C.; Oli e , G.; Lesko, J.; Ba es, K.; e al.
COVID-19 mo ali y p edic ion om deep lea ning in a la ge mul is a e EHR and LIS da ase : Algo i hm de elopmen and
alida ion (P ep in ). J. Med. In e ne Res. 2021,23, e30157. [C ossRe ]
97.
Schöning, V.; Liakoni, E.; Baumga ne , C.; Exadak ylos, A.; Hau z, W.; A kinson, A.; Hammann, F. De elopmen and alida ion
o a p ognos ic COVID-19 se e i y assessmen (COSA) sco e and machine lea ning models o pa ien iage a a e ia y hospi al.
J. T ansl. Med. 2021,19, 56. [C ossRe ]
98.
Zhu, J.; Ge, P.; Jiang, C.; Zhang, Y.; Li, X.; Zhao, Z.; Zhang, L.; Duong, T. Deep-lea ning a i icial in elligence analysis o clinical
a iables p edic s mo ali y in COVID-19 pa ien s. J. Am. Coll. Eme g. Physicians Open 2020,1, 1364–1373. [C ossRe ] [PubMed]
99.
Li, S.; Lin, Y.; Zhu, T.; Fan, M.; Xu, S.; Qiu, W.; Chen, C.; Li, L.; Wang, Y.; Yan, J.; e al. De elopmen and ex e nal e alua ion
o p edic ions models o mo ali y o COVID-19 pa ien s using machine lea ning me hod. Neu al Compu . Appl. 2021,35,
13037–13046. [C ossRe ] [PubMed]
100.
Qui oz, J.C.; Feng, Y.Z.; Cheng, Z.Y.; Rezazadegan, D.; Chen, P.K.; Lin, Q.T.; Qian, L.; Liu, X.F.; Be ko sky, S.; Coie a, E.; e al.
De elopmen and Valida ion o a Machine Lea ning App oach o Au oma ed Se e i y Assessmen o COVID-19 Based on
Clinical and Imaging Da a: Re ospec i e S udy. JMIR Med. In o m. 2021,9, e24572. [C ossRe ] [PubMed]
101.
de Fá ima, A.; S emel, D.; Fachi, M.; Su ek, M.; Wiens, A.; S ump Tonin, F.; Pon a olo, R. Diagnosis and p edic ion o COVID-19
se e i y: Can biochemical es s and machine lea ning be used as p ognos ic indica o s? Compu . Biol. Med. 2021,134, 104531.
[C ossRe ]
102.
Boo h, A.; Abels, E.; McCa ey, P. De elopmen o a p ognos ic model o mo ali y in COVID-19 in ec ion using machine
lea ning. Mod. Pa hol. 2020,34, 522–531. [C ossRe ] [PubMed]
103.
Zhao, Y.; Chen, Q.; Liu, T.; Luo, P.; Zhou, Y.; Liu, M.; Xiong, B.; Zhou, F. De elopmen and Valida ion o P edic o s o he Su i al
o Pa ien s wi h COVID-19 Based on Machine Lea ning. F on . Med. 2021,8, 683431. [C ossRe ]
104.
Hu, C.; Liu, Z.; Jiang, Y.; Shi, O.; Zhang, X.; Xu, K.; Suo, C.; Wang, Q.; Song, Y.; Yu, K.; e al. Ea ly p edic ion o mo ali y isk
among pa ien s wi h se e e COVID-19, using machine lea ning. In . J. Epidemiol. 2020,49, 1918–1929. [C ossRe ]
105.
Zhou, K.; Ss, T.; Li, L.; Zang, Z.; Wang, J.; Li, J.; Liang, J.; Zhang, F.; Zhang, Q.; Ge, W.; e al. Ele en Rou ine Clinical Fea u es
P edic COVID-19 Se e i y Unco e ed by Machine Lea ning o Longi udinal Measu emen s. Compu . S uc . Bio echnol. J. 2021,
19, 3640–3649. [C ossRe ]
106.
Shi, W.; Peng, X.; Liu, T.; Cheng, Z.; Lu, H.; Yang, S.; Zhang, J.; Wang, M.; Gao, Y.; Shi, Y.; e al. A deep lea ning-based quan i a i e
compu ed omog aphy model o p edic ing he se e i y o COVID-19: A e ospec i e s udy o 196 pa ien s. Ann. T ansl. Med.
2021,9, 216. [C ossRe ] [PubMed]
107.
Casi aghi, E.; Malchiodi, D.; T ucco, G.; F asca, M.; Cappelle i, L.; Fon ana, T.; Esposi o, A.; A ola, E.; Jache i, A.; Reese, J.; e al.
Explainable Machine Lea ning o Ea ly Assessmen o COVID-19 Risk P edic ion in Eme gency Depa men s. IEEE Access 2020,
8, 196299–196325. [C ossRe ] [PubMed]
108.
Ye, J.; Hua, M.; Zhu, F. Machine Lea ning Algo i hms a e Supe io o Con en ional Reg ession Models in P edic ing Risk
S a i ica ion o COVID-19 Pa ien s. Risk Manag. Heal hc. Policy 2021,14, 3159–3166. [C ossRe ]
109.
Azna -Gimeno, R.; Es eban, L.; Lezaun, G.; del Hoyo-Alonso, R.; Abadia-Gallego, D.; Paño-Pa do, J.; Esquillo -Rod igo, M.;
Lanas, A.; Se ano, M.T. A Clinical Decision Web o P edic ICU Admission o Dea h o Pa ien s Hospi alised wi h COVID-19
Using Machine Lea ning Algo i hms. In . J. En i on. Res. Public Heal h 2021,18, 8677. [C ossRe ]
110.
Dabbah, M.; Reed, A.; Boo h, A.; Yassaee, A.; Despo o ic, A.; Klasme , B.; Binning, E.; A al, M.; Plans, D.; Mo elli, D.; e al.
Machine lea ning app oach o dynamic isk modeling o mo ali y in COVID-19: A UK Biobank s udy. Sci. Rep. 2021,11, 16936.
[C ossRe ] [PubMed]
111.
Ko, H.; Chung, H.; Kang, W.S.; Pa k, C.; Kim, D.W.; Kim, S.E.; Chung, C.R.; Ko, R.E.; Lee, H.; Seo, J.H.; e al. An A i icial
In elligence Model o P edic he Mo ali y o COVID-19 Pa ien s a Hospi al Admission Time Using Rou ine Blood Samples:
De elopmen and Valida ion o an Ensemble Model. J. Med. In e ne Res. 2020,22, e25442. [C ossRe ]
112.
Sánchez-Mon añés, M.; Rod iguez, P.; Se ano-López, A.; Oli as, E.; Alakhda -Mohma a, Y. Machine Lea ning o Mo ali y
Analysis in Pa ien s wi h COVID-19. In . J. En i on. Res. Public Heal h 2020,17, 8386. [C ossRe ] [PubMed]
113.
Pa el, D.; Khe , V.; Desai, B.; Lei, X.; Cen, S.; Nanda, N.; Gholam ezanezhad, A.; Duddalwa , V.; Va ghese, B.; Obe ai, A. Machine
lea ning based p edic o s o COVID-19 disease se e i y. Sci. Rep. 2021,11, 4673. [C ossRe ]

Elec onics 2024,13, 1005 33 o 38
114.
Ak a , S.; Talukde , A.; Ahamad, M.; Kamal, A.; Khan, J.; Hossain, N.; Azad, A.K.M.; Quinn, J.; Summe s, M.; Liaw, S.T.; e al.
Machine Lea ning App oaches o Iden i y Pa ien Como bidi ies and Symp oms Tha Inc eased Risk o Mo ali y in COVID-19.
Diagnos ics 2021,11, 1383. [C ossRe ] [PubMed]
115.
Gao, Y.; Cai, G.; Fang, W.; Li, H.Y.; Wang, S.Y.; Chen, L.; Yu, Y.; Liu, D.; Xu, S.; Cui, P.F.; e al. Machine Lea ning Based Ea ly
Wa ning Sys em Enables Accu a e Mo ali y Risk P edic ion o COVID-19. Na . Commun. 2020,11, 5033. [C ossRe ]
116.
Yu, L.; Halalau, A.; Dalal, B.; Abbas, A.; I ascu, F.; Amin, M.; Nai , G. Machine lea ning me hods o p edic mechanical en ila ion
and mo ali y in pa ien s wi h COVID-19. PLoS ONE 2021,16, e0249285. [C ossRe ]
117.
Shi i, I.; Salimi, Y.; Pakbin, M.; Hajian a , G.; Haddadi A al, A.; Sanaa , A.; Mos a aei, S.; Akha analla , A.; Sabe i Manesh, A.;
Mansou i, Z.; e al. COVID-19 p ognos ic modeling using CT adiomic ea u es and machine lea ning algo i hms: Analysis o a
mul i-ins i u ional da ase o 14,339 pa ien s. Compu . Biol. Med. 2022,145, 105467. [C ossRe ]
118.
Lip ák, P.; Bano cin, P.; Rosol’anka, R.; P okopiˇc, M.; Kocan, I.; Žiaˇciko á, I.; Uh ik, P.; G endá , M.; Hy del, R. A machine
lea ning app oach o iden i ica ion o gas oin es inal p edic o s o he isk o COVID-19 ela ed hospi aliza ion. Pee J 2022,
10, e13124. [C ossRe ]
119.
Ya bakhsh, R.; Mo aza i, S.; Mo aza i, S.; Pa saei, H.; Rad, D. A i icial in elligence e ec i ely p edic s he COVID-19 dea h
a e in di e en UK ci ies. J. In ell. Fuzzy Sys . 2022,43, 1853–1857. [C ossRe ]
120.
Mohammad, R.; Aljab i, M.; Aboulnou , M.; Mi za, S.; Alshobaiki, A. Classi ying he Mo ali y o People wi h Unde lying Heal h
Condi ions A ec ed by COVID-19 Using Machine Lea ning Techniques. Appl. Compu . In ell. So Compu . 2022,2022, 3783058.
[C ossRe ]
121.
Baik, S.M.; Lee, M.; Hong, K.S.; Pa k, D.J. De elopmen o Machine-Lea ning Model o P edic COVID-19 Mo ali y: Applica ion
o Ensemble Model and Rega ding Fea u e Impac s. Diagnos ics 2022,12, 1464. [C ossRe ] [PubMed]
122.
Wan, T.K.; Huang, R.X.; Tulu, T.; Liu, J.D.; Vodenca e ic, A.; Wong, C.W.; Chan, K.h. Iden i ying P edic o s o COVID-19 Mo ali y
Using Machine Lea ning. Li e 2022,12, 547. [C ossRe ] [PubMed]
123.
Azizi, Z.; Shiba, Y.; Alipou , P.; Maleki, F.; Rapa elli, V.; No is, C.; Fo ghani, R.; Pilo e, L.; Emam, K. Impo ance o sex and
gende ac o s o COVID-19 in ec ion and hospi alisa ion: A sex-s a i ied analysis using machine lea ning in UK Biobank da a.
BMJ Open 2022,12, e050450. [C ossRe ] [PubMed]
124.
Vezzoli, M.; Incia di, R.; O iecuia, C.; Pa is, S.; Mu illo, N.; Agos oni, P.; Ame i, P.; Bellasi, A.; Campo o ondo, R.; Canale, C.; e al.
Machine lea ning o p edic ion o in-hospi al mo ali y in co ona i us disease 2019 pa ien s: Resul s om an I alian mul icen e
s udy. J. Ca dio asc. Med. 2022,23, 439–446. [C ossRe ] [PubMed]
125.
Reina Reina, A.; Ba e a, J.; Valdi ieso, B.; Gas Lopez, M.E.; Ma é, A.; T ujillo, J. Machine lea ning model om a Spanish coho
o p edic ion o SARS-CoV-2 mo ali y isk and c i ical pa ien s. Sci. Rep. 2022,12, 5723. [C ossRe ] [PubMed]
126.
Rana, A.; Singh, H.; Ma udu u, R.; Pa anaik, S.; Rana, P. Quan i ying p ognosis se e i y o COVID-19 pa ien s om deep
lea ning based analysis o CT ches images. Mul imed. Tools Appl. 2022,81, 18129–18153. [C ossRe ]
127.
Chen, L.; Mei, Z.; Guo, W.; Ding, S.; Huang, T.; Cai, Y.D. Recogni ion o Immune Cell Ma ke s o COVID-19 Se e i y wi h
Machine Lea ning Me hods. BioMed Res. In . 2022,2022, 6089242. [C ossRe ]
128.
Mazloumi, R.; Abaza i, S.R.; Na a ieh, F.; Aghsami, A.; Jolai, F. S a is ical analysis o blood cha ac e is ics o COVID-19
pa ien sand hei su i al o dea h p edic ion using machine lea ningalgo i hms. Neu al Compu . Appl. 2022,34, 14729–14743.
[C ossRe ]
129.
Ag awal, S.; Pa il, N. Machine Lea ning based COVID-19 Mo ali y P edic ion using Common Pa ien Da a. In P oceedings
o he 2022 IEEE 7 h In e na ional Con e ence o Con e gence in Technology (I2CT), Mumbai, India, 7–9 Ap il 2022; pp. 1–6.
[C ossRe ]
130.
Laino, M.; Gene ali, E.; Tommasini, T.; Angelo i, G.; Aghemo, A.; Desai, A.; Mo andini, P.; S e anini, G.; Lleo, A.; Voza, A.; e al.
An Indi idualized Algo i hm o P edic Mo ali y in COVID-19 Pneumonia: A Machine Lea ning Based S udy. A ch. Med. Sci.
2022,18, 587. [C ossRe ] [PubMed]
131.
Be am, M.G.; Sundin, J.; Roche, D.G.; Sánchez-Tója , A.; Tho é, E.S.J.; B odin, T. Open science. Cu . Biol. 2023,33, R792–R797.
[C ossRe ] [PubMed]
132.
Ali, O.; Ishak, M.K.; Bha i, M.K.L. A Machine Lea ning App oach o Ea ly COVID-19 Symp oms Iden i ica ion. Compu . Ma e .
Con in. 2022,70, 3803–3820. [C ossRe ]
133.
Tadepalli, S.; Thulasi am, R. COVID-19 Ea ly Symp om P edic ion Using Blockchain and Machine Lea ning. In Blockchain and
Applica ions: 3 d In e na ional Cong ess; Sp inge : Be lin/Heidelbe g, Ge many, 2022; pp. 243–251. [C ossRe ]
134.
Sil a, C.; Junio , A.; Lopes, R. P edic i e Analysis o COVID-19 Symp oms in Social Ne wo ks h ough Machine Lea ning.
Elec onics 2022,11, 580. [C ossRe ]
135.
Chen, Z.; Li, M.; Wang, R.; Sun, W.; Liu, J.; Li, H.; Wang, T.; Lian, Y.; Zhang, J.; Wang, X. Diagnosis o COVID-19 ia Acous ic
Analysis and A i icial In elligence by Moni o ing B ea h Sounds on Sma phones. J. Biomed. In o m. 2022,130, 104078. [C ossRe ]
136.
Lee, D.; Wang, C.; McAlis e , F.; Ma, S.; Chu, A.; Rochon, P.; Kaul, P.; Aus in, P.; Wang, X.; Kalmady, S.; e al. Fac o s associa ed
wi h SARS-CoV-2 es posi i i y in long- e m ca e homes: A popula ion-based coho analysis using machine lea ning. Lance
Reg. Heal h Am. 2022,6, 100146. [C ossRe ] [PubMed]
137.
Go ji, F.; Sha iekhani, S.; Namda , P.; Abdollahzade, S.; Ra iei, S. Machine lea ning-based COVID-19 diagnosis by demog aphic
cha ac e is ics and clinical da a. Ad . Respi . Med. 2022,90, 171–183. [C ossRe ]
Elec onics 2024,13, 1005 34 o 38
138.
Sha ma, D.; Sub amanian, M.; Malyad i, P.; Reddy, B.; Sha ma, D.; Tah eem, M. Classi ica ion o COVID -19 by Using Supe ised
Op imized Machine Lea ning Technique. Ma e . Today P oc. 2021,56, 2058–2062. [C ossRe ]
139.
Chadaga, K.; Chak abo y, C.; P abhu, S.; Umakan h, S.; Bha , V.; Sampa hila, N. Clinical and Labo a o y App oach o Diagnose
COVID-19 Using Machine Lea ning. In e discip. Sci. Compu . Li e Sci. 2022,14, 452–470. [C ossRe ]
140.
Thimo eo, L.; Vellasco, M.; Ama al, J.; Figuei edo, K.; Yokoyama, C.; Ma ques, E. Explainable A i icial In elligence o COVID-19
Diagnosis Th ough Blood Tes Va iables. J. Con ol Au om. Elec . Sys . 2022,33, 625–644. [C ossRe ]
141.
S idha , A.; Chen, Z.H.; May ield, J.; Fohne , A.; A ani is, P.; A kinson, S.; B aunschweig, F.; Cha e jee, N.; Zamponi, A.;
Johnson, G.; e al.
Iden i ying Risk o Ad e se Ou comes in COVID-19 Pa ien s ia A i icial In elligence-Powe ed Analysis o
12-Lead In ake Elec oca diog am. Ca dio asc. Digi . Heal h J. 2021,3, 62–74. [C ossRe ] [PubMed]
142.
Kapoo , A.; Kapoo , A.; Mahajan, G.; Kapu , A. Use o A i icial In elligence o T iage Pa ien s wi h Flu-Like Symp oms Using
Imaging in Non- COVID-19 Hospi als du ing COVID-19 Pandemic: An Ongoing 8-Mon h Expe ience. Indian J. Radiol. Imaging
2022,31, 901–909. [C ossRe ] [PubMed]
143.
Cab as, S. A Bayesian-Deep Lea ning Model o Es ima ing COVID-19 E olu ion in Spain. Ma hema ics 2021,9, 2921. [C ossRe ]
144.
Sol an, A.; Yang, J.; Pa anshe y, R.; No ak, A.; Yang, Y.; Rohanian, O.; Bee , S.; Sol an, M.; Thicke , D.; Fai head, R.; e al.
Real-wo ld e alua ion o apid and labo a o y- ee COVID-19 iage o eme gency ca e: Ex e nal alida ion and pilo deploymen
o a i icial in elligence d i en sc eening. Lance Digi . Heal h 2022,4, e266–e278. [C ossRe ]
145.
Kim, H.J.; Han, D.; Kim, J.H.; Kim, D.; Ha, B.; Seog, W.; Lee, Y.K.; Lim, D.; Hong, S.; Pa k, M.J.; e al. An Easy- o-Use Machine
Lea ning Model o P edic he P ognosis o Pa ien s Wi h COVID-19: Re ospec i e Coho S udy. J. Med. In e ne Res. 2020,
22, e24225. [C ossRe ]
146.
Muelle , Y.; Sch ama, T.; Ruij en, R.; Sch eu s, M.; G asho , D.; an de We ken, H.; Jona Lasinio, G.; Al a ez-de la Sie a,
D.; Kie nan, C.; Ei o, M.; e al. S a i ica ion o hospi alized COVID-19 pa ien s in o clinical se e i y p og ession g oups by
immuno-pheno yping and machine lea ning. Na . Commun. 2022,13, 915. [C ossRe ]
147.
Hou, W.; Zhao, Z.; Chen, A.; Li, H.; Duong, T. Machining lea ning p edic s he need o escala ed ca e and mo ali y in COVID-19
pa ien s om clinical a iables. In . J. Med. Sci. 2021,18, 1739–1745. [C ossRe ]
148.
Campbell, T.W.; Wilson, M.P.; Rode , H.; MaWhinney, S.; Geo gan as, R.W.; Magui e, L.K.; Rode , J.; E landson, K.M. P edic ing
p ognosis in COVID-19 pa ien s using machine lea ning and eadily a ailable clinical da a. In . J. Med. In o m. 2021,155, 104594.
[C ossRe ]
149.
Vaid, A.; Somani, S.; Russak, A.; F ei as, J.; Chaudh y, F.; Pa anjpe, I.; Johnson, K.; Lee, S.; Mio o, R.; Rich e , F.; e al. Machine
Lea ning o P edic Mo ali y and C i ical E en s in a Coho o Pa ien s Wi h COVID-19 in New Yo k Ci y: Model De elopmen
and Valida ion. J. Med. In e ne Res. 2020,22, e24018. [C ossRe ]
150.
Ka , S.; Chawla, R.; Ha ana h, S.; Ramasubban, S.; Ramak ishnan, N.; Vaishya, R.; Sibal, A.; Reddy, S. Mul i a iable mo ali y isk
p edic ion using machine lea ning o COVID-19 pa ien s a admission (AICOVID). Sci. Rep. 2021,11, 12801. [C ossRe ]
151.
Xu, W.; Sun, N.N.; Gao, H.N.; Chen, Z.Y.; Yang, Y.; Bin, J.; Tang, L.L. Risk ac o s analysis o COVID-19 pa ien s wi h ARDS and
p edic ion based on machine lea ning. Sci. Rep. 2021,11, 2933. [C ossRe ]
152.
Rahman, T.; Khandaka , A.; Haque, M.; Ib ehaz, N.; Kashem, S.; Islam, M.; Al-Maadeed, S.; Zughaie , S.; Doi, S.; Chowdhu y, M.
De elopmen and Valida ion o an Ea ly Sco ing Sys em o P edic ion o Disease Se e i y in COVID-19 Using Comple e Blood
Coun Pa ame e s. IEEE Access 2021,9, 120422–120441. [C ossRe ] [PubMed]
153.
Chung, H.; Ko, H.; Kang, W.S.; Kim, K.W.; Lee, H.; Pa k, C.; Song, H.O.; Choi, T.Y.; Seo, J.H.; Lee, J. P edic ion and Fea u e
Impo ance Analysis o Se e i y o COVID-19 in Sou h Ko ea Using A i icial In elligence: Model De elopmen and Valida ion.
J. Med. In e ne Res. 2021,23, e27060. [C ossRe ] [PubMed]
154.
Rahman, T.; Al-Ishaq, F.; Al-Mohannadi, F.; Muba ak, R.; Al-Hi mi, M.; Islam, K.; Khandaka , A.; Ai Hssain, A.; Al-Madeed, S.;
Zughaie , S.; e al. Mo ali y P edic ion U ilizing Blood Bioma ke s o P edic he Se e i y o COVID-19 Using Machine Lea ning
Technique. Diagnos ics 2021,11, 1582. [C ossRe ] [PubMed]
155.
Ikemu a, K.; Golds ein, D.; Szymanski, J.; Bellin, E.; S ahl, L.; Yagi, Y.; Saada, M.; Simone, K.; Reyes, M.; Bille , H. Using
Au oma ed-Machine Lea ning o P edic COVID-19 Pa ien Mo ali y (P ep in ). J. Med. In e ne Res. 2020,23, e23458. [C ossRe ]
[PubMed]
156.
Mune a, N.; Ga cia-Gallo, E.; Gonzalez, Á.; Zea, J.; Fuen es, Y.; Se ano, C.; Ruiz-Cua as, A.; Rod íguez, A.; Reyes, L. A no el
model o p edic se e e COVID-19 and mo ali y using an a i icial in elligence algo i hm o in e p e ches X-Rays and clinical
a iables. ERJ Open Res. 2022,8, 00010–2022. [C ossRe ] [PubMed]
157.
Cihan, P. The machine lea ning app oach o p edic ing he numbe o in ensi e ca e, in uba ed pa ien s and dea h: The COVID-19
pandemic in Tu key. Sigma J. Eng. Na . Sci. 2022,40, 85–94. [C ossRe ]
158.
Izquie do, J.L.; Ancochea, J.; So iano, J.B. Clinical Cha ac e is ics and P ognos ic Fac o s o In ensi e Ca e Uni Admission o
Pa ien s Wi h COVID-19: Re ospec i e S udy Using Machine Lea ning and Na u al Language P ocessing. J. Med. In e ne Res.
2020,22, e21801. [C ossRe ] [PubMed]
159.
Nino, G.; Lingu a u, M.G. De eloping a i icial in elligence echnology o pedia ic pulmonology: Lessons om COVID-19.
Pedia . Pulmonol. 2022,57, 1588. [C ossRe ]
160.
Subudhi, S.; Ve ma, A.; Pa el, A.; Ha din, C.; Khandeka , M.; Lee, H.; Mce oy, D.; S ylianopoulos, T.; Munn, L.; Du a, S.; e al.
Compa ing machine lea ning algo i hms o p edic ing ICU admission and mo ali y in COVID-19. npj Digi . Med. 2021,4, 87.
[C ossRe ]
Elec onics 2024,13, 1005 35 o 38
161. Shamou , F.; Shen, Y.; Wu, N.; Kaku, A.; Pa k, J.; Makino, T.; Jas z˛ebski, S.; Wi owski, J.; Wang, D.; Zhang, B.; e al. An a i icial
in elligence sys em o p edic ing he de e io a ion o COVID-19 pa ien s in he eme gency depa men . npj Digi . Med. 2021,
4, 80. [C ossRe ]
162.
A é alo, J.; Gómez, J.; Casas, J.; An ón-San os, J.; Mele o-Be mejo, J.; López-Ca mona, M.; Palacios, L.; Sanz-Cáno as, J.;
Pesquei a-Fon án, P.; Peña-Fe nández, A.; e al. The impo ance o associa ion o como bidi ies on COVID-19 ou comes: A
machine lea ning app oach. Cu . Med. Res. Opin. 2022,38, 501–510. [C ossRe ]
163.
Kalaba ige, L.R.; Ma ingan i, H. Symp om Based COVID-19 Tes Recommenda ion Sys em Using Machine Lea ning Technique.
In ell. Decis. Technol. 2022,16, 181–191. [C ossRe ]
164.
Li, X.; Ge, P.; Zhu, J.; Li, H.; G aham, J.; Singe , A.; Richman, P.; Duong, T. Deep lea ning p edic ion o likelihood o ICU admission
and mo ali y in COVID-19 pa ien s using clinical a iables. Pee J 2020,8, e10337. [C ossRe ]
165.
A ind, V.; Kim, J.; Cho, B.; Geng, E.; Cho, S. De elopmen o a machine lea ning algo i hm o p edic in uba ion among
hospi alized pa ien s wi h COVID-19. J. C i . Ca e 2021,62, 25–30. [C ossRe ]
166.
Famiglini, L.; Campagne , A.; Ca obene, A.; Cabi za, F. A obus and pa simonious machine lea ning me hod o p edic ICU
admission o COVID-19 pa ien s. Med. Biol. Eng. Compu . 2022. [C ossRe ] [PubMed]
167.
Bolou ani, S.; B enne , M.; Wang, P.; McGinn, T.; Hi sch, J.S.; Ba naby, D.; Zanos, T.P. A Machine Lea ning P edic ion Model o
Respi a o y Failu e Wi hin 48 Hou s o Pa ien Admission o COVID-19: Model De elopmen and Valida ion. J. Med. In e ne
Res. 2021,23, e24246. [C ossRe ]
168.
Izadi, Z.; Gian ancesco, M.; Hy ich, K.; S ang eld, A.; Gossec, L.; Ca mona, L.; Ma eus, E.; Lawson-To ey, S.; T upin, L.;
Rush, S.; e al.
Machine lea ning algo i hms o p edic COVID-19 acu e espi a o y dis ess synd ome in pa ien s wi h heuma ic
diseases: Resul s om he global heuma ology alliance p o ide egis y. Ann. Rheum. Dis. 2021,80, 175–176. [C ossRe ]
169. Bouhamed, H.; Hamdi, M.; Ga gou i, R. COVID-19 Pa ien s’ Hospi al Occupancy P edic ion Du ing he Recen Omic on Wa e
ia some Recu en Deep Lea ning A chi ec u es. In . J. Compu . Commun. Con ol 2022,17, 4697. [C ossRe ]
170.
Laino, M.; Ammi abile, A.; Lo ino, L.; Lundon, D.; Chi i, A.; F ancone, M.; Sa e ski, V. P ognos ic indings o ICU admission in
pa ien s wi h COVID-19 pneumonia: Baseline and ollow-up ches CT and he added alue o a i icial in elligence. Eme g. Radiol.
2022,29, 243–262. [C ossRe ] [PubMed]
171.
Lopes, F.P.P.; Ki amu a, F.; P ado, G.; Ku iki, P.; Ga cia, M. Machine lea ning model o p edic ing se e i y p ognosis in pa ien s
in ec ed wi h COVID-19: S udy p o ocol om COVID-AI B asil. PLoS ONE 2021,16, e0245384. [C ossRe ]
172.
an de Leu , R.; Bleijendaal, H.; Taha, K.; Mas , T.; Gho, J.; Linscho en, M.; Rees, B.; Henkens, M.; Heymans, S.;
S u kenboom, N.; e al.
Elec oca diog am-based mo ali y p edic ion in pa ien s wi h COVID-19 using machine lea n-
ing. Ne h. Hea J. 2022,30, 312–318. [C ossRe ] [PubMed]
173.
Ma ia, A.; Dimi ios, V.; Ioanna, M.; Cha alampos, M.; Ge asimos, M.; Cons an inos, K. Clinical Decision Making and Ou come
P edic ion o COVID-19 Pa ien s Using Machine Lea ning. In Pe asi e Compu ing Technologies o Heal hca e, P oceedings o he
15 h EAI In e na ional Con e ence, Pe asi e Heal h 2021, Vi ual E en , 6–8 Decembe 2021; Lewy, H., Ba kan, R., Eds.; Sp inge :
Cham, Swi ze land, 2022; pp. 3–14.
174.
Huang, F.; Chen, L.; Guo, W.; Zhou, X.; Feng, K.; Huang, T.; Cai, Y. Iden i ying COVID-19 Se e i y-Rela ed SARS-CoV-2 Mu a ion
Using a Machine Lea ning Me hod. Li e 2022,12, 806. [C ossRe ] [PubMed]
175.
Boddu, R.S.K.; Ka maka , P.; Bhaumik, A.; Nassa, V.K.; Vandana; Bha acha ya, S. Analyzing he impac o Machine lea ning and
A i icial in elligence and i s E ec on Managemen o lung cance de ec ion in COVID-19 pandemic. Ma e . Today P oc. 2021,56,
2213–2216. [C ossRe ] [PubMed]
176.
Pa el, N.; D’Sil a, K.; Li, M.; Hsu, T.; DiIo io, M.; Fu, X.; Cook, C.; P isco, L.; Ma in, L.; Vanni, K.; e al. Assessing he Se e i y o
COVID-19 Lung Inju y in Rheuma ic Diseases e sus he Gene al Popula ion Using Deep Lea ning-De i ed Ches Radiog aph
Sco es. A h i is Ca e Res. 2022,75, 657–666. [C ossRe ]
177.
Gadipudi, P.; Teja, P.; Yelamancheli, C.; Thanise ika an, A.; Mohiuddin, R.; Joy, A. De ec ion o pneumonia p og ession in lungs
o indi iduals a ec ed wi h COVID-19 se e ely using deep lea ning echniques. In P oceedings o he 2022 3 d In e na ional
Con e ence o Eme ging Technology (INCET), Belgaum, India, 27–29 May 2022; pp. 1–5. [C ossRe ]
178.
Zhang, X.; Lu, S.; Wang, S.; Yu, X.; Wang, S.J.; Yao, L.; Pan, Y.; Zhang, Y.D. Diagnosis o COVID-19 pneumonia ia a no el deep
lea ning a chi ec u e. J. Compu . Sci. Technol. 2022,37, 0679. [C ossRe ]
179.
A ab, M.; Amin, R.; Koundal, D.; Aldabbas, H.; Alou i, B.; Iqbal, Z. Classi ica ion o COVID-19 and In luenza Pa ien s Using
Deep Lea ning. Con as Media Mol. Imaging 2022,2022, 8549707. [C ossRe ]
180.
Jingxin, L.; Mengchao, Z.; Yuchen, L.; Jinglei, C.; Yu ong, Z.; Zhong, Z.; Lihui, Z. COVID-19 lesion de ec ion and segmen a ion—A
deep lea ning me hod. Me hods 2021,202, 62–69. [C ossRe ]
181.
Shahin, O.; M. Abd El Aziz, R.; Taloba, A. De ec ion and classi ica ion o COVID-19 in CT-lungs sc eening using machine lea ning
echniques. J. In e discip. Ma h. 2022,25, 791–813. [C ossRe ]
182.
Aswa hy, A.; Anand, H.; Chand a, V. COVID-19 se e i y de ec ion using machine lea ning echniques om CT-images. E ol.
In ell. 2022,16, 1423–1431. [C ossRe ]
183.
Shahin, O.; Alshamma i, H.; Taloba, A.; M. Abd El Aziz, R. Machine Lea ning App oach o Au onomous De ec ion and
Classi ica ion o COVID-19 Vi us. Compu . Elec . Eng. 2022,101, 108055. [C ossRe ]
184.
Chie ega o, M.; F angiamo e, F.; Mo assi, M.; Ba esi, C.; Nici, S.; Basse i, C.; Bnà, C.; Galelli, M. A hyb id machine lea ning/deep
lea ning COVID-19 se e i y p edic i e model om CT images and clinical da a. Sci. Rep. 2022,12, 4329. [C ossRe ]
Elec onics 2024,13, 1005 36 o 38
185.
Venka a amana, L.; P asad, D.; Shunmugana han, S.; Mi huma y, C.; Ka hikeyan, R.; Monika, N. Classi ica ion o COVID-19
om ube culosis and pneumonia using deep lea ning echniques. Med. Biol. Eng. Compu . 2022,60, 2681–2691. [C ossRe ]
[PubMed]
186.
Aswa hy, A.L.; Ha eend an, A.; Vinod Chand a, S.S. COVID-19 diagnosis and se e i y de ec ion om CT-images using ans e
lea ning and back p opaga ion neu al ne wo k. J. In ec . Public Heal h 2021,14, 1435–1445. [C ossRe ]
187.
Pu kayas ha, S.; Xiao, Y.; Jiao, Z.; Thepumnoeysuk, R.; Halsey, K.; Wu, J.; T an, L.; Hsieh, B.; Choi, J.W.; Wang, D.; e al. Machine
Lea ning-Based P edic ion o COVID-19 Se e i y and P og ession o C i ical Illness Using CT Imaging and Clinical Da a. Ko ean
J. Radiol. 2021,22, 1213. [C ossRe ]
188.
Gazzah, S.; Bayi, R.; Kaloun, S.; Bencha e , O. A Deep Lea ning o Dis inguish COVID-19 om O he s Pneumonia Cases. In ell.
Au om. So Compu . 2021,31, 677–693. [C ossRe ]
189. Elkamouny, M.; Ghan ous, M. Pneumonia Classi ica ion o COVID-19 Based on Machine Lea ning. In P oceedings o he 2022
2nd In e na ional Mobile, In elligen and Ubiqui ous Compu ing Con e ence (MIUCC), Cai o, Egyp , 8–9 May 2022; pp. 135–140.
[C ossRe ]
190.
Liu, Q.; Pang, B.; Li, H.; Zhang, B.; Liu, Y.; Lai, L.; Le, W.; Li, J.; Xia, T.; Zhang, X.; e al. Machine lea ning models o p edic ing
c i ical illness isk in hospi alized pa ien s wi h COVID-19 pneumonia. J. Tho ac. Dis. 2021,13, 1215. [C ossRe ] [PubMed]
191.
Ma, H.; Ye, Q.; Ding, W.; Jiang, Y.; Wang, M.; Niu, Z.; Zhou, X.; Gao, Y.; Wang, C.; Menpes-Smi h, W.; e al. Can Clinical Symp oms
and Labo a o y Resul s P edic CT Abno mali y? Ini ial Findings Using No el Machine Lea ning Techniques in Child en Wi h
COVID-19 In ec ions. F on . Med. 2021,8, 699984. [C ossRe ] [PubMed]
192.
Lu, X.; Cui, Z.; Pan, F.; Li, L.; Li, L.; Liang, B.; Yang, L.; Zheng, C. Glycemic s a us a ec s he se e i y o co ona i us disease
2019 in pa ien s wi h diabe es melli us: An obse a ional s udy o CT adiological mani es a ions using an a i icial in elligence
algo i hm. Ac a Diabe ol. 2021,58, 575–586. [C ossRe ] [PubMed]
193.
Vepa, A.; Saleem, A.; Rakhshan, K.; Daneshkhah, A.; Sedighi, T.; Shohaimi, S.; Oma , A.; Sala i, N.; Cha abgoun, O.;
Dha ma aj, D.; e al.
Using Machine Lea ning Algo i hms o De elop a Clinical Decision-Making Tool o COVID-19 Inpa-
ien s. In . J. En i on. Res. Public Heal h 2021,18, 6228. [C ossRe ]
194.
Guzmán-To es, J.A.; Alonso-Guzmán, E.M.; Domínguez-Mo a, F.J.; Tinoco-Gue e o, G. Es ima ion o he main condi ions in
(SARS-CoV-2) COVID-19 pa ien s ha inc ease he isk o dea h using Machine lea ning, he case o Mexico. Resul s Phys. 2021,
27, 104483. [C ossRe ]
195. Iacola e, B.; Pe one, V.; Sangio gi, D.; Ghigi, A.; Giacomini, E.; Nappi, C.; Paoli, D.; Ancona, D.; And e a, M.; Ba bie i, A.; e al.
POSA310 A i icial In elligence Applied on Adminis a i e Big Da a o P edic he Se e i y o SARS-CoV-2 In ec ion. Value Heal h
2022,25, S199–S200. [C ossRe ]
196. Ho, D. Add essing COVID-19 D ug De elopmen wi h A i icial In elligence. Ad . In ell. Sys . 2020,2, 2000070. [C ossRe ]
197.
A o a, G.; Joshi, J.; Mandal, R.; Sh i as a a, N.; Vi mani, R.; Se hi, T. A i icial In elligence in Su eillance, Diagnosis, D ug
Disco e y and Vaccine De elopmen agains COVID-19. Pa hogens 2021,10, 1048. [C ossRe ] [PubMed]
198.
Schul z, M.; Ve a, D.; Sinclai , D. Can a i icial in elligence iden i y e ec i e COVID-19 he apies? Embo Mol. Med. 2020,
12, e12817. [C ossRe ] [PubMed]
199.
Zhang, Y.; Ye, T.; Xi, H.; Juhas, M.; Li, J. Deep Lea ning D i en D ug Disco e y: Tackling Se e e Acu e Respi a o y Synd ome
Co ona i us 2. F on . Mic obiol. 2021,12, 739684. [C ossRe ] [PubMed]
200.
We ne , J.; K onbe g, R.; S achu a, P.; Os e mann, P.; Mülle , L.; Schaal, H.; Bha ia, S.; Ka he , J.; Bo kha d , A.; Pandy a, A.; e al.
Deep T ans e Lea ning App oach o Au oma ic Recogni ion o D ug Toxici y and Inhibi ion o SARS-CoV-2. Vi uses 2021,
13, 610. [C ossRe ]
201.
Majumda , S.; Nandi, S.; Ghosal, S.; Ghosh, B.; Mallik, W.; Du a Roy, N.; Biswas, A.; Mukhe jee, S.; Pal, S.; Bha acha yya, N.
Deep Lea ning-Based Po en ial Ligand P edic ion F amewo k o COVID-19 wi h D ug–Ta ge In e ac ion Model. Cogn. Compu .
2021. [C ossRe ]
202.
Mikkili, I.; Peele, A.; Veka eswa ulu, T.; Vidya P abhaka , K.; Macamdas, D.; S ee ama, K. Po en ial o a i icial in elligence o
accele a e diagnosis and d ug disco e y o COVID-19. Pee J 2021,9, e12073. [C ossRe ]
203.
Gaw iljuk, V.; Zin, P.P.; Puhl, A.; Zo n, K.; Foil, D.; Lane, T.; Hu s , B.; Almeida Ta ella, T.; Cos a, F.;
Lakshmanane, P.; e al.
Machine Lea ning Models Iden i y Inhibi o s o SARS-CoV-2. J. Chem. In . Model. 2021,61, 4224–4235. [C ossRe ]
204.
Su a na, K.; Biswas, D.; Pai, M.G.; Acha jee, A.; Banka , R.; Palani el, V.; Salka , A.; Ve ma, A.; Mukhe jee, A.;
Choudhu y, M.; e al.
P o eomics and Machine Lea ning App oaches Re eal a Se o P ognosi c Ma ke s o COVID-19
Se e i y Wi h D ug Repu posing Po en ial. F on . Physiol. 2021,12, 652799. [C ossRe ]
205.
Han, L.; Shan, G.; Chu, B.; Wang, H.; Wang, Z.; Gao, S.; Zhou, W. Accele a ing d ug epu posing o COVID-19 ea men by
modeling mechanisms o ac ion using cell image ea u es and machine lea ning. Cogn. Neu odyn. 2021,17, 803–811. [C ossRe ]
206.
Aghdam, R.; Habibi, M.; Tahe i, G. Using in o ma i e ea u es in machine lea ning based me hod o COVID-19 d ug epu posing.
J. Chemin o m. 2021,13, 70. [C ossRe ] [PubMed]
207.
Jha, N.; P asha , D.; Rashid, M.; Sha iq, M.; Khan, R.; P uncu, C.; Siddiqui, S.; Sa a ana Kuma , M. Deep Lea ning App oach o
Disco e y o in Silico D ugs o Comba ing COVID-19. J. Heal hc. Eng. 2021,2021, 6668985. [C ossRe ] [PubMed]
208.
Pham, T.H.; Qiu, Y.; Zeng, J.; Xie, L.; Zhang, P. A deep lea ning amewo k o high- h oughpu mechanism-d i en pheno ype
compound sc eening and i s applica ion o COVID-19 d ug epu posing. Na . Mach. In ell. 2021,3, 247–257. [C ossRe ] [PubMed]
Elec onics 2024,13, 1005 37 o 38
209.
Auwul, M.; Rahman, M.R.; Go , E.; Shahjaman, M.; Moni, M.A. Bioin o ma ics and machine lea ning app oach iden i ies
po en ial d ug a ge s and pa hways in COVID-19. B ie ings Bioin o m. 2021,22, bbab120. [C ossRe ] [PubMed]
210.
Bha i, A.; Wan, S.; Al è, D.; Clyde, A.; Bode, M.; Tan, L.; Ti o , M.; Me zky, A.; Tu illi, M.; Jha, S.; e al. Pandemic D ugs a
Pandemic Speed: In as uc u e o Accele a ing COVID-19 D ug Disco e y wi h Hyb id Machine Lea ning- and Physics-based
Simula ions on High Pe o mance Compu e s. In e ace Focus 2021,11, 20210018. [C ossRe ]
211.
Abdel-Basse , M.; Hawash, H.; Elhoseny, M.; Chak abo y, R.; Ryan, M. DeepH-DTA: Deep Lea ning o P edic ing D ug-Ta ge
In e ac ions: A Case S udy o COVID-19 D ug Repu posing. IEEE Access 2020,8, 170433–170451. [C ossRe ]
212.
Nguyen, T.; Pham, D.H.; Hiep, D.; Phuong-Thao, T.; Quang, D.; Ngo, S.T. Sea ching and designing po en ial inhibi o s o
SARS-CoV-2 Mp o om na u al sou ces using a omis ic and deep-lea ning calcula ions. RSC Ad . 2021,11, 38495–38504.
[C ossRe ]
213.
Zhou, Y.; Wang, F.; Tang, J.; Nussino , R.; Cheng, F. A i icial in elligence in COVID-19 d ug epu posing. Lance . Digi . Heal h
2020,2, e667–e676. [C ossRe ]
214.
Koli z, S.; Kim, J.; Zhang, J.; Cha, Y.; Ba ula, S.; Kusko, R.; K ishnan, R.; Zeskind, B.; Kau man, H. 477 Deep lea ning o d i e
COVID-19 apid d ug epu posing. J. Immuno he . Cance 2020,8, A509. [C ossRe ]
215.
Louce a, C.; Es eban-Medina, M.; Rian, K.; Ma ín Falco, M.; Dopazo, J.; Peña-Chile , M. D ug epu posing o COVID-19 using
machine lea ning and mechanis ic models o signal ansduc ion ci cui s ela ed o SARS-CoV-2 in ec ion. Signal T ansduc . Ta ge .
The . 2020,5, 290. [C ossRe ] [PubMed]
216.
Mohapa a, S.; Na h, P.; Cha e jee, M.; Das, N.; Kali a, D.; Roy, P.; Sa apa hi, S. Repu posing he apeu ics o COVID-19: Rapid
p edic ion o comme cially a ailable d ugs h ough machine lea ning and docking. PLoS ONE 2020,15, e0241543. [C ossRe ]
[PubMed]
217.
El-Behe y, H.; A ia, A.F.; El-Fishawy, N.; To key, H. E icien machine lea ning model o p edic ing d ug- a ge in e ac ions wi h
case s udy o COVID-19. Compu . Biol. Chem. 2021,93, 107536. [C ossRe ] [PubMed]
218.
Rajpu , A.; Thaku , A.; Mukhopadhyay, A.; Kamboj, S.; Ras ogi, A.; Gau am, S.; Jassal, H.; Kuma , M. P edic ion o epu posed
d ugs o Co ona i uses using a i icial in elligence and machine lea ning. Compu . S uc . Bio echnol. J. 2021,19, 3133–3148.
[C ossRe ] [PubMed]
219.
Jin, W.; S okes, J.; Eas man, R.; I kin, Z.; Zakha o , A.; Collins, J.; Jaakkola, T.; Ba zilay, R. Deep lea ning iden i ies syne gis ic
d ug combina ions o ea ing COVID-19. P oc. Na l. Acad. Sci. USA 2021,118, e2105070118. [C ossRe ] [PubMed]
220.
Bu dick, H.; Lam, C.; Ma a aso, S.; Sie kas, A.; B aden, G.; Dellinge , R.; McCoy, A.; Vincen , J.L.; G een-Saxena, A.;
Ba nes, G.; e al.
Is Machine Lea ning a Be e Way o Iden i y COVID-19 Pa ien s Who Migh Bene i om Hyd oxychlo oquine
T ea men ?—The IDENTIFY T ial. J. Clin. Med. 2020,9, 3834. [C ossRe ] [PubMed]
221.
Zeng, X.; Song, X.; Ma, T.; Pan, Q.; Zhou, Y.; Hou, Y.; Zhang, Z.; Li, K.; Ka ypis, G.; Cheng, F. Repu pose Open Da a o Disco e
The apeu ics o COVID-19 using Deep Lea ning. J. P o eome Res. 2020,19, 4624–4636. [C ossRe ]
222.
Su, X.; Hu, L.; You, Z.; Hu, P.; Wang, L.; Zhao, B. A deep lea ning me hod o epu posing an i i al d ugs agains new i uses ia
mul i- iew nonnega i e ma ix ac o iza ion and i s applica ion o SARS-CoV-2. B ie ings Bioin o m. 2021,23, bbab526. [C ossRe ]
223.
San os, S.; To es, M.; Sánchez, M.; Ce nuzzi, L.; Paccana o, A. Machine Lea ning and Ne wo k Medicine app oaches o D ug
Reposi ioning o COVID-19. Pa e ns 2021,3, 100396. [C ossRe ]
224.
Beck, B.R.; Shin, B.; Choi, Y.; Pa k, S.; Kang, K. P edic ing comme cially a ailable an i i al d ugs ha may ac on he no el
co ona i us (SARS-CoV-2) h ough a d ug- a ge in e ac ion deep lea ning model. Compu . S uc . Bio echnol. J. 2020,18, 784–790.
[C ossRe ] [PubMed]
225.
Bung, N.; K ishnan, S.; Bulusu, G.; Roy, A. De no o design o new chemical en i ies o SARS-CoV-2 using a i icial in elligence.
Fu u e Med. Chem. 2021,13, 575–585. [C ossRe ] [PubMed]
226.
Ke, Y.Y.; Peng, T.T.; Yeh, T.K.; Huang, W.Z.; Chang, S.E.; Wu, S.H.; Hung, H.C.; Hsu, T.A.; Lee, S.J.; Song, J.S.; e al. A i icial
in elligence app oach igh ing COVID-19 wi h epu posing d ugs. Biomed. J. 2020,43, 355–362. [C ossRe ] [PubMed]
227.
Yang, W.; Li, Q.; Sun, J.; Tan, S.; Tang, Y.H.; Zhao, M.; Li, Y.Y.; Cao, X.; Zhao, J.C.; Yang, J.K. Po en ial D ug Disco e y o
COVID-19 T ea men Ta ge ing Ca hepsin L Using a Deep Lea ning-Based S a egy. Compu . S uc . Bio echnol. J. 2022,20,
2442–2454. [C ossRe ] [PubMed]
228.
Mekni, N.; Co onnello, C.; Lange , T.; De Rosa, M.; Pe icone, U. Suppo Vec o Machine as a Supe ised Lea ning o he
P io i iza ion o No el Po en ial SARS-CoV-2 Main P o ease Inhibi o s. In . J. Mol. Sci. 2021,22, 7714. [C ossRe ] [PubMed]
229.
Kowalewski, J.; Ray, A. P edic ing no el d ugs o SARS-CoV-2 using machine lea ning om a >10 million chemical space.
Heliyon 2020,6, e04639. [C ossRe ]
230.
Pin o, G.P.; Va a, O.; Ma ques, S.M.; Filipo ic, J.; Bedna , D.; Dambo sky, J. Sc eening o wo ld app o ed d ugs agains highly
dynamical spike glycop o ein o SARS-CoV-2 using Ca e Dock and machine lea ning. Compu . S uc . Bio echnol. J. 2021,
19, 3187–3197. [C ossRe ] [PubMed]
231.
Zame, W.; Bica, I.; Shen, C.; Cu h, A.; Lee, H.S.; Bailey, S.; Wea he all, J.; W igh , D.; B e z, F.; Schaa , M. Machine Lea ning o
Clinical T ials in he E a o COVID-19. S a . Biopha m. Res. 2020,12, 506–517. [C ossRe ]
232.
da Mo a, O.J.; Sil a, E.; Siquei a-Ba is a, R. Bioe hical aspec s o a i icial in elligence: COVID-19 & end o li e. Re . Assoc. Méd.
B as. 2020,66, 5–6. [C ossRe ]

Elec onics 2024,13, 1005 38 o 38
233.
Jung, C.; Exco ie , J.B.; Raphaël-Rousseau, M.; Salaün-Penque , N.; O ala, M.; Chouaid, C. E olu ion o hospi alized pa ien
cha ac e is ics h ough he i s h ee COVID-19 wa es in Pa is a ea using machine lea ning analysis. PLoS ONE 2022,17, e0263266.
[C ossRe ] [PubMed]
234.
Asi , S.; Saba, T.; Alghanim, A. Explo ing P edic ion o COVID-19 and i s Se e i y using Machine Lea ning. In P oceedings o he
2022 Fi h In e na ional Con e ence o Women in Da a Science a P ince Sul an Uni e si y (WiDS PSU), Riyadh, Saudi A abia,
28–29 Ma ch 2022; pp. 117–122. [C ossRe ]
235.
Co es, U.; Co es, A.; Ga cia-Gasulla, D.; Pe ez-A nal, R.; A a ez-Napagao, S.; Al a ez, E. The e hical use o high-pe o mance
compu ing and a i icial in elligence: Figh ing COVID-19 a Ba celona Supe compu ing Cen e . AI E hics 2022,2, 325–340.
[C ossRe ] [PubMed]
Disclaime /Publishe ’s No e: The s a emen s, opinions and da a con ained in all publica ions a e solely hose o he indi idual
au ho (s) and con ibu o (s) and no o MDPI and/o he edi o (s). MDPI and/o he edi o (s) disclaim esponsibili y o any inju y o
people o p ope y esul ing om any ideas, me hods, ins uc ions o p oduc s e e ed o in he con en .