Maddodi, S i a sa; Kun e, S ini asa Rao
A icle
Ma ke esilience in u bulen imes: A p oac i e app oach
o p edic ing s ock ma ke esponses du ing geopoli ical
ensions
Jou nal o Capi al Ma ke s S udies (JCMS)
P o ided in Coope a ion wi h:
Tu kish Capi al Ma ke s Associa ion
Sugges ed Ci a ion: Maddodi, S i a sa; Kun e, S ini asa Rao (2024) : Ma ke esilience in u bulen
imes: A p oac i e app oach o p edic ing s ock ma ke esponses du ing geopoli ical ensions,
Jou nal o Capi al Ma ke s S udies (JCMS), ISSN 2514-4774, Eme ald, Bingley, Vol. 8, Iss. 2, pp.
173-194,
h ps://doi.o g/10.1108/JCMS-12-2023-0049
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/313319
S anda d-Nu zungsbedingungen:
Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen
Zwecken und zum P i a geb auch gespeiche und kopie we den.
Sie dü en die Dokumen e nich ü ö en liche ode komme zielle
Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich
machen, e eiben ode ande wei ig nu zen.
So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen
(insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en,
gel en abweichend on diesen Nu zungsbedingungen die in de do
genann en Lizenz gewäh en Nu zungs ech e.
Te ms o use:
Documen s in EconS o may be sa ed and copied o you pe sonal
and schola ly pu poses.
You a e no o copy documen s o public o comme cial pu poses, o
exhibi he documen s publicly, o make hem publicly a ailable on he
in e ne , o o dis ibu e o o he wise use he documen s in public.
I he documen s ha e been made a ailable unde an Open Con en
Licence (especially C ea i e Commons Licences), you may exe cise
u he usage igh s as speci ied in he indica ed licence.
h ps://c ea i ecommons.o g/licenses/by/4.0/
Ma ke esilience in u bulen
imes: a p oac i e app oach o
p edic ing s ock ma ke esponses
du ing geopoli ical ensions
S i a sa Maddodi
Depa men o Compu e Science, S ini as Uni e si y, Mangalu u, India and
Depa men o Da a Enginee ing, Uni ed Ai lines Inc, Bengalu u, India, and
S ini asa Rao Kun e
Depa men o Compu e Science, S ini as Uni e si y, Mangalu u, India
Abs ac
Pu pose –The Indian s ock ma ke can be icky when he e’s ouble in he wo ld, like wa s o big con lic s.
I ’s like ying o ead a sec e message. We wan o igu e ou wha makes in es o s ne ous o happy,
because hei eelings o en a ec how hey buy and sell s ocks. We’ e building a ool o make p edic ion ha
uses bo h numbe s and people’s opinions.
Design/me hodology/app oach –Hyb id app oach le e ages Twi e sen imen , ma ke da a, ola ili y
index (VIX) and momen um indica o s like mo ing a e age con e gence di e gence (MACD) and ela i e
s eng h index (RSI) o deli e accu a e ma ke insigh s o in o med in es men decisions du ing unce ain y.
Findings –Ou s udy e eals ha geopoli ical ensions’ impac on s ock ma ke s is lee ing and con ined o
he sho e m. Capi alizing on his insigh , we buil a g ound-b eaking p edic i e model wi h an imp essi e
98.47% accu acy in o ecas ing s ock ma ke alues du ing such e en s.
O iginali y/ alue –To he bes o he au ho s’ knowledge, his model’s o iginali y lies in i s ocus on sho -
e m impac , no el da a usion and high accu acy. Focus on sho - e m impac : Ou model uniquely iden i ies
and quan i ies he lee ing e ec s o geopoli ical ensions on ma ke beha io , a p e iously unde - esea ched
a ea. No el da a usion: Combining sen imen analysis wi h es ablished ma ke indica o s like VIX and
momen um o e s a comp ehensi e and dynamic app oach o p edic ing ma ke mo emen s du ing ola ile
pe iods. Ad anced p edic i e accu acy: Achie ing he p edic ion accu acy (98.47%) se s his model apa
om exis ing solu ions, making i a aluable ool o in o med decision-making.
Keywo ds Geopoli ical ension, S ock ma ke p edic ion, Sen imen analysis, Momen um indica o , LSTM
Pape ype Technical pape
1. In oduc ion
The e icien ma ke hypo hesis (EMH) sugges s ha s ock p ices encapsula e all a ailable
in o ma ion, inco po a ing his o ical p ices and exchange-based sha e ades, ading a ai
alues conside ing all ele an ac o s (Fama, 1970,1991). Howe e , schola ly esea ch o e
he yea s has aised ques ions abou whe he s ock p ice luc ua ions may also be in luenced
by he beha io al and a ional aspec s o in es o s (Shlei e and Summe s, 1990). These
elemen s, collec i ely known as “in es o sen imen ” o he “mood o he in es o ,” in oduce
a human elemen in o he ma ke dynamics. The ise o social media pla o ms such as
Twi e , Facebook, YouTube and LinkedIn has p o ided an ins an aneous digi al ou le o
Jou nal o Capi al
Ma ke s S udies
173
© S i a sa Maddodi and S ini asa Rao Kun e. Published in Jou nal o Capi al Ma ke s S udies.
Published by Eme ald Publishing Limi ed. This a icle is published unde he C ea i e Commons
A ibu ion (CC BY 4.0) license. Anyone may ep oduce, dis ibu e, ansla e and c ea e de i a i e
wo ks o his a icle ( o bo h comme cial and non-comme cial pu poses), subjec o ull a ibu ion o
he o iginal publica ion and au ho s. The ull e ms o his license may be seen a h p://
c ea i ecommons.o g/licences/by/4.0/legalcode
The cu en issue and ull ex a chi e o his jou nal is a ailable on Eme ald Insigh a :
h ps://www.eme ald.com/insigh /2514-4774.h m
Recei ed 12 Feb ua y 2024
Re ised 27 July 2024
18 Augus 2024
Accep ed 2 Sep embe 2024
Jou nal o Capi al Ma ke s S udies
Vol. 8 No. 2, 2024
pp. 173-194
Eme ald Publishing Limi ed
2514-4774
DOI 10.1108/JCMS-12-2023-0049
indi iduals o exp ess opinions and eac ions in ex ual o m. Resea che s can employ
app op ia e echniques o analyze hese ex s and disce n he sen imen behind public
exp essions on hese pla o ms. Geopoli ics encompasses a b oad spec um o e en s wi h
di e se causes and consequences, anging om e o is a acks o clima e change and om
B exi o he Global Financial C isis (Bekae e al., 2014). Such e en s and h ea s in oduce
unce ain ies and isks in o inancial ma ke s, including he s ock exchange, gi ing ise o
“geopoli ical isk” (GPR), a signi ican conce n o s akeholde s such as policymake s,
co po a ions, inancial in es o s and poli icians. GPR plays a pi o al ole in shaping in es o
decisions ega ding s ock ansac ions. Unce ain y, as emphasized by Campbell e al., is a
undamen al aspec in inancial economics, in luencing in es o beha io and ma ke p ices
(Be k and DeMa zo, 2017). GPRs encompass ac o s like “geopoli ical ensions,” “wa isk,”
“ e o is h ea s” and “ ade wa s be ween coun ies.” His o ical e idence highligh s he
nega i e impac o e en s such as he Sep embe 11, 2001, e o is a ack and he I aq
in asion on global s ock exchanges. The annexa ion o C imea and Russia’s sanc ions in
2014 also esul ed in signi ican d ops in Russian s ock indices. Fo ecas ing s ock ma ke
beha io du ing geopoli ical ensions p esen s a o midable challenge due o he ola ile
na u e o he s ock ma ke . While news and his o ical p ices a e commonly conside ed
in luencing ac o s, ecen esea ch emphasizes he s ong link be ween in es o sen imen
and s ock ma ke e u ns in bo h he USA and Eu opean ma ke s (Bake and Wu gle , 2000;
Das and Vasileios, 2014;Lee and Wu, 2016;Shehzad and Malik, 2019). This pape aims o
sc u inize he Indian s ock ma ke du ing pe iods o GPR, speci ically ocusing on ensions
be ween India and Pakis an ollowing he URI e o is a ack in Pulwama and India’s
subsequen su gical s ike on e o is camps. Addi ionally, we examine he ongoing
geopoli ical ension be ween Russia and Uk aine, gi en India’s close ies wi h Russia.
Employing analy ical app oaches, ou objec i e is o p o ide imp o ed p edic ions and a
deepe unde s anding o he beha io o he Indian s ock ma ke du ing hese c i ical
pe iods. The analysis concen a es on he Na ional S ock Exchange (NSE), one o India’s
majo s ock exchanges, u ilizing indexes like he Ni y Nex 50 Index and he Ni y Midcap
100 Index, wi h he lagship Ni y 50 Index comp ising 50 companies widely u ilized by bo h
Indian and global in es o s as a ba ome e o he Indian equi y ma ke .
2. Rela ed wo k
This li e a u e e iew is di ided in o wo main sec ions:
Li e a u e ela ed o he subs an i e issue: This sec ion examines exis ing esea ch on how
geopoli ical ensions impac s ock ma ke s and he po en ial o p edic ion.
Li e a u e ela ed o he heo e ical amewo k: This sec ion explo es ele an heo ies
om beha io al inance and ne wo k heo y ha can in o m ou unde s anding o how social
media sen imen and ma ke indica o s migh in luence in es o beha io du ing
geopoli ical e en s.
2.1 Li e a u e ela ed o he subs an i e issue
Plakanda as e al. (2018) explo e he use o machine lea ning, speci ically suppo ec o
eg ession (SVR), o p edic inancial ma ke beha io du ing pe iods o geopoli ical
unce ain y. Thei s udy e eals ha SVR o e s imp o ed o ecas ing accu acy o e
adi ional models, especially du ing ma ke u bulence. Howe e , he e ec i eness o SVR
a ies ac oss di e en ime ames and sec o s, wi h gold ma ke s showing he s onges
esponse. The au ho s no e ha ce ain sec o s, such as echnology and ou ism, may eac
di e en ly o geopoli ical e en s. They emphasize he need o u he esea ch o e ine hese
models and add ess limi a ions o b oade applicabili y.
JCMS
8,2
174
Alqah ani e al. (2020) assess he impac o GPR and c ude oil e u ns on s ock ma ke
p edic abili y in six Gul Coope a ion Council (GCC) coun ies, using da a om Feb ua y
2007 o Decembe 2019. They ind ha while GPR indices show limi ed abili y o p edic
s ock e u ns wi hin he sample, he global GPR index has some ou -o -sample o ecas ing
alue o Kuwai and Oman. C ude oil p ices gene ally p o ide s onge p edic i e powe o
GCC s ock e u ns, bo h wi hin and beyond he sample pe iod. These indings emain obus
e en when accoun ing o isk-adjus ed e u ns. The s udy sugges s ha while GPR
in luences GCC s ock ma ke s, c ude oil p ices should be he p ima y ocus o in es o s
seeking sho - e m o ecas s and inco po a ing addi ional a iables could enhance u u e
p edic ion models.
Salisu e al. (2022a,b) examine he impac o global GPR on he s ock ma ke s o
ad anced economies, including he G7 and Swi ze land, o e mo e han a cen u y. They ind
ha GPR signi ican ly p edic s s ock e u ns, wi h he an icipa ion o geopoli ical e en s
o en causing mo e ma ke dis up ion han he e en s hemsel es. Gold p o es o be a
eliable hedge du ing geopoli ical u bulence, showing a nega i e co ela ion wi h s ock
ma ke ola ili y. The s udy also no es ha eme ging ma ke s a e mo e sensi i e o
geopoli ical h ea s, especially when egional ade in eg a ion is in ol ed. The esea ch
unde sco es he need o conside GPR sen imen , no jus e en s, o mo e accu a e ma ke
p edic ions and sugges s ha policymake s could use his unde s anding o manage isks
and enhance s abili y.
T iki and Maa oug (2021) in es iga e he ela ionship be ween he USA s ock ma ke and
gold p ices du ing pe iods o geopoli ical ension and con lic . They ind ha gold nega i ely
co ela es wi h he S anda d & Poo ’s 500 (S&P 500) Index, se ing as a hedge agains s ock
ma ke ola ili y, especially when GPR is high. The s udy in oduces he GPR index as an
e ec i e measu e o global poli ical ensions, showing i s impac on ma ke s. The co ela ion
be ween gold and he s ock ma ke is dynamic, sugges ing he need o adap able models in
esponse o changing geopoli ical condi ions. While he s udy ocuses on he USA ma ke , i s
indings may be ele an o o he egions wi h simila dynamics, and u he esea ch is
ecommended o explo e gold’s ole as a sa e ha en ac oss di e en con ex s.
Segnon e al. (2023) examine he e ec i eness o using GPR indica o s o p edic s ock
ma ke ola ili y. Using a obus au o eg essi e-mul i a ia e-s ochas ic gene alized
au o eg essi e condi ional he e oskedas ici y mixed da a sampling (AR-MSGARCH-
MIDAS) model ha accoun s o s uc u al b eaks and bo h sho - and long- e m
ola ili y, hey ind ha including GPR a iables does no signi ican ly enhance he
accu acy o mon hly ola ili y o ecas s. The s udy sugges s ha he impac o GPR on
o ecas ing may depend on he speci ic p edic ion model used. A e accoun ing o
non-s a iona i ies, he added in o ma ion om GPR was no s a is ically signi ican , hough
mac oeconomic a iables migh o e complemen a y insigh s in ce ain cases. Despi e using
o e 122 yea s o da a, he s udy highligh s he challenge o di ec ly cap u ing he p edic i e
powe o geopoli ical e en s on ma ke ola ili y, sugges ing ha u he esea ch could
explo e al e na i e GPR da a o model e inemen s.
Salisu e al. (2022a,b) examine he GPRs in eme ging ma ke s, inding ha s ock ma ke
ola ili y in hese ma ke s inc eases in esponse o geopoli ical h ea s, pa icula ly he
an icipa ion o e en s a he han he e en s hemsel es. The s udy highligh s he po en ial
o machine lea ning, especially SVR, in o ecas ing ola ili y du ing geopoli ical u bulence,
wi h gold being pa icula ly sensi i e. I emphasizes he impo ance o ocusing on
geopoli ical sen imen o be e ma ke p edic abili y and no es ha eme ging ma ke s a e
mo e ulne able o GPR due o ade in eg a ion. The indings sugges ha di e en sec o s
eac di e en ly o GPR, wi h long- e m implica ions o in es o con idence and economic
g ow h, unde sco ing he need o policymake s o unde s and he GPR–ma ke link o
manage isks e ec i ely.
Jou nal o Capi al
Ma ke s S udies
175
Fio illo e al. (2023) examine he impac o GPR on s ock liquidi y in in e na ional ma ke s,
inding ha highe GPR leads o dec eased liquidi y, making i ha de o buy and sell sha es
quickly and a low cos . The s udy e eals ha he an icipa ion and h ea o geopoli ical
e en s ha e a mo e signi ican impac on liquidi y han he ac ual e en s hemsel es. S ocks
in less liquid ma ke s o hose issued by inancially cons ained o less- anspa en
companies a e pa icula ly ulne able o he nega i e e ec s o GPR. These indings sugges
ha GPR impac s liquidi y h ough inc eased inancing cons ain s and in o ma ion
asymme y. The esea ch highligh s he impo ance o unde s anding he GPR–liquidi y
link o bo h in es o s and policymake s.
Hachicha (2023) examined he GPR du ing he Russia–Uk aine wa , highligh ing a s ong
in e dependence be ween in es o sen imen , exchange a es, GPR and de eloping s ock
ma ke s ac oss a ious ime ho izons. The s udy inds ha in es o sen imen o en leads o
changes in exchange a es, pa icula ly in he sho e m, and bo h in es o sen imen and
GPR signi ican ly in luence s ock ma ke e u ns, especially o e he long e m. Gold is
iden i ied as an e ec i e hedge agains cu ency de alua ion and s ock ma ke ins abili y
due o i s nega i e co ela ion wi h hese a iables. Eme ging ma ke s a e pa icula ly
sensi i e o luc ua ions in GPR and in es o sen imen , in ensi ying he obse ed
co-mo emen s. The s udy unde sco es he impo ance o conside ing bo h in es o
sen imen and GPR when analyzing de eloping s ock ma ke s du ing pe iods o
geopoli ical unce ain y.
Heds €
om e al. (2020) in es iga e he impac o geopoli ical unce ain y on s ock ma ke
con agion in eme ging ma ke s, inding ha hese ma ke s a e mo e ulne able o con agion
wi hin hei egions han om global ma ke s, sugges ing po en ial di e si ica ion bene i s
by in es ing ac oss di e en egions. The s udy e eals ha gene al s ock ma ke isk
[ ola ili y index (VIX)] is a mo e signi ican d i e o ma ke ola ili y han GPR indices,
which do no s ongly impac e u n o ola ili y spillo e s be ween ma ke s. S ong ade
in eg a ion wi hin egional ma ke s con ibu es o he heigh ened egional con agion.
Addi ionally, he s udy con i ms a nega i e co ela ion be ween gold p ices and s ock
ma ke ola ili y du ing geopoli ical unce ain y. O e all, he indings emphasize he
impo ance o unde s anding egional ade ne wo ks and gene al ma ke isks in managing
con agion in eme ging ma ke s.
2.2 Li e a u e ela ed o he heo e ical amewo k
This s udy u ilizes a combined lens om beha io al inance and ne wo k heo y o
unde s and how social media sen imen and ma ke indica o s can in luence s ock ma ke
mo emen s du ing geopoli ical ensions.
Beha io al inance: P ospec heo y by Kahneman and T e sky (1979) sugges s ha
in es o s exhibi loss a e sion, po en ially leading o isk a e sion and selling du ing
pe iods o unce ain y e lec ed in nega i e social media sen imen . Addi ionally, he d
beha io migh be ampli ied by social media, leading o collec i e o e eac ions du ing
geopoli ical e en s.
Ne wo k heo y: The amewo k o complex adap i e sys ems can be applied o
unde s and he s ock ma ke as a ne wo k whe e indi idual in es o ac ions based on social
media sen imen can ha e unp edic able collec i e e ec s on ma ke beha io . This aligns
wi h he wo k on ma ke e iciency by Fama (1970), whe e semi-s ong e iciency sugges s
ha public in o ma ion (including social media) migh no be ully p iced in o he ma ke ,
o e ing an oppo uni y o p edic ion.
Bollen e al. (2011) analyzed Twi e mood and disco e ed ha i can p edic s ock ma ke
mo emen s. They ound a co ela ion be ween collec i e mood s a es on social media and
ma ke ends. The s udy sugges s ha social media sen imen analysis can be a use ul ool
JCMS
8,2
176
o o ecas ing ma ke beha io , as in es o s’ emo ional esponses o geopoli ical e en s
e lec ed on social media can impac ma ke dynamics.
Ba be is (2013) e iewed he impac o p ospec heo y on economics o e 3 decades,
no ing i s wide applicabili y in explaining di e se economic beha io s. The key indings
emphasize ha he p ospec heo y o e s a c ucial amewo k o unde s anding in es o
beha io . In eg a ing beha io al insigh s om his heo y can enhance inancial models and
imp o e ma ke p edic ions.
Shille (2000) in es iga ed he psychological ac o s in luencing ma ke dynamics,
ocusing on he oles o i a ional exube ance and he d beha io in c ea ing asse bubbles.
The key indings highligh ha he d beha io can lead o signi ican ma ke o e eac ions
and ha social media can ampli y hese e ec s by apidly and widely sp eading
sen imen s.
Hi shlei e and Teoh (2003) p o ided a comp ehensi e e iew o he ding beha io in
capi al ma ke s, ocusing on how in o ma ion cascades and social lea ning con ibu e o his
phenomenon. Thei key indings emphasize ha he d beha io is a signi ican d i e o
ma ke ola ili y. Unde s anding he in luence o social in o ma ion can imp o e he abili y
o p edic and in e p e ma ke mo emen s.
Bane jee (1992) de eloped a model o explain he d beha io in economics, demons a ing
how indi iduals o en imi a e o he s’ ac ions when hey belie e o he s possess be e
in o ma ion. The key indings e eal ha he d beha io can esul in subop imal decision-
making and ha du ing geopoli ical c ises, ma ke pa icipan s may ollow p e ailing ends
a he han making independen assessmen s.
Ahe n and Sosyu a (2015) examined he impac o sensa ionalis media co e age on s ock
p ices, inding ha such news signi ican ly a ec s s ock e u ns and ola ili y. Thei key
indings highligh ha media, including social media, plays a c ucial ole in shaping in es o
pe cep ions and ma ke eac ions. Unde s anding his media in luence is i al o p edic ing
ma ke beha io , especially du ing geopoli ical e en s.
Wa s (2002) de eloped a model o explain global cascades on andom ne wo ks,
demons a ing how small ini ial shocks can igge widesp ead e ec s h oughou a
ne wo k. The key indings sugges ha inancial ma ke s can unde go signi ican changes
om mino e en s due o ne wo k e ec s, and social media can ac as a ca alys o such
cascades, pa icula ly du ing pe iods o geopoli ical unce ain y.
Te lock (2007) analyzed how media con en in luences in es o sen imen and ma ke
ou comes, inding ha a nega i e media one ends o p edic downwa d p essu e on s ock
p ices. The key akeaways a e ha media sen imen analysis is essen ial o an icipa ing
ma ke mo emen s and ha in es o s’ esponses o geopoli ical news can be signi ican ly
shaped by he media’s one and aming.
Cookson e al. (2020) in es iga ed how social media ampli ies poli ical ex emism,
e ealing ha echo chambe s con ibu e o pola ized opinions and beha io . Thei indings
highligh ha social media can enhance in es o biases and lead o ex eme ma ke eac ions.
Moni o ing social media sen imen is, he e o e, c ucial o p edic ing ma ke esponses,
pa icula ly du ing geopoli ical c ises.
Li e al. (2019) examined he impac o social media on he inancial pe o mance o ini ial
public o e ings (IPOs), inding ha posi i e social media sen imen signi ican ly boos s IPO
pe o mance. The key akeaways a e ha a o able social media sen imen can enhance
in es o con idence and imp o e ma ke pe o mance. Addi ionally, analyzing social media
ends can o e aluable insigh s in o ma ke dynamics, especially du ing geopoli ical
e en s.
Lough an and McDonald (2011) de eloped a me hod o ex ual analysis o inancial
disclosu es, iden i ying ha speci ic wo ds and ph ases co ela e wi h ma ke pe o mance.
Thei key indings sugges ha ex ual analysis o social media can be a powe ul ool o
Jou nal o Capi al
Ma ke s S udies
177
p edic ing ma ke mo emen s. Unde s anding he language used in social media pos s can
help gauge in es o sen imen and an icipa e po en ial ma ke eac ions.
2.3 Resea ch gap
While a subs an ial amoun o esea ch explo es he link be ween geopoli ical ensions and
s ock ma ke mo emen s, exis ing s udies o en ely solely on public sen imen analysis om
social media. The li e a u e highligh s se e al a eas o u he in es iga ion:
(1) Limi ed in eg a ion o sen imen analysis and ma ke indica o s: While sen imen
analysis and adi ional indica o s like he VIX and momen um indica o s (MACD
and RSI) ha e been explo ed sepa a ely, he e is a lack o s udies in eg a ing hese
app oaches o enhance s ock ma ke p edic ions du ing geopoli ical e en s.
(2) Dynamic and mul i ace ed models: Many s udies ocus on s a ic models o single
indica o s. The e is a need o dynamic, mul i ace ed models ha adap o changing
geopoli ical condi ions and in eg a e a ious ypes o da a (e.g. in es o sen imen ,
public sen imen and adi ional indica o s).
To add ess hese gaps, we p opose a mo e comp ehensi e app oach ha inco po a es bo h
in es o and public sen imen analysis, alongside adi ional ma ke indica o s like he VIX
and momen um indica o s (MACD, RSI, e c.). This mul i ace ed app oach o e s dis inc
ad an ages o s ock ma ke p edic ion:
2.4 Enhanced accu acy h ough a mul i-laye ed lens
(1) Gauging ma ke mood: Sen imen analysis helps cap u e he emo ional s a e o
in es o s and he gene al public, which can signi ican ly in luence ma ke mo emen s.
Op imis ic sen imen may lead o buying sp ees, while ea can igge sello s.
(2) Cap u ing b oade ends and in es o insigh s: In es o sen imen analysis, in
pa icula , can e eal in es o s’ pe cep ions o speci ic companies o indus ies,
po en ially impac ing hei s ock p ices.
(1) Imp o ing p edic ion capabili ies:
•Iden i ying ea ly ends: By analyzing sen imen ends alongside adi ional
indica o s like he VIX, in es o s can po en ially iden i y po en ial shi s in ma ke
sen imen be o e hey a e ully e lec ed in s ock p ices.
•Mi iga ing isk: Du ing pe iods o high ola ili y ( e lec ed by he VIX),
unde s anding in es o sen imen can help in es o s make in o med decisions and
po en ially educe isk.
2.5 Le e aging he powe o adi ional indica o s
(1) VIX as a ola ili y gauge: The VIX, o en e e ed o as he “ ea gauge,” p o ides
aluable insigh s in o ma ke sen imen du ing pe iods o unce ain y. In eg a ing
hese da a alongside sen imen analysis can o e a mo e comple e unde s anding o
ma ke psychology.
(2) Momen um indica o s o ma ke dynamics: Momen um indica o s like MACD and
RSI can cap u e he di ec ion and s eng h o p ice mo emen s. This in o ma ion,
combined wi h sen imen analysis, can help p edic sho - e m ma ke ends,
pa icula ly du ing geopoli ical e en s.
JCMS
8,2
178
2.6 Limi a ions and he impo ance o a balanced app oach
(1) Ma ke noise: Sen imen analysis can be in luenced by i ele an in o ma ion online.
Fil e ing and in e p e ing he da a accu a ely is c ucial.
(2) Sel - ul illing p ophecies: Nega i e sen imen can po en ially lead o ma ke
down u ns, c ea ing a sel - ul illing p ophecy. I ’s impo an o conside his
possibili y when using sen imen analysis.
O e all, analyzing bo h in es o and public sen imen alongside adi ional ma ke
indica o s o e s aluable insigh s o s ock ma ke o ecas ing, pa icula ly du ing
pe iods o geopoli ical ins abili y. This comp ehensi e app oach can po en ially imp o e he
accu acy o sho - e m p edic ion models by p o iding a deepe unde s anding o he
psychological and social ac o s in luencing in es o beha io .
2.7 Resea ch objec i es and ques ions
This esea ch aims o de elop a no el app oach o he Indian s ock ma ke :
2.8 Objec i es
(1) De elop a hyb id model combining social media sen imen analysis wi h ma ke
indica o s;
(2) Assess he sho - e m impac o geopoli ical ensions on he Indian s ock ma ke and
(3) E alua e he model’s e ec i eness in p edic ing s ock ma ke mo emen s du ing
such e en s.
2.9 Resea ch ques ions
RQ1. Can social media sen imen analysis iden i y in es o ne ousness o con idence
du ing geopoli ical ensions?
RQ2. How e ec i ely do adi ional ma ke indica o s (VIX, MACD and RSI) cap u e he
ma ke ’s esponse o geopoli ical e en s?
RQ3. How accu a ely can a hyb id model p edic sho - e m s ock ma ke mo emen s
du ing geopoli ical ensions?
2.10 Hypo hesis
A hyb id model inco po a ing social media sen imen analysis wi h adi ional ma ke
indica o s will ou pe o m exis ing models in p edic ing sho - e m s ock ma ke mo emen s
du ing geopoli ical ensions.
3. Me hodology
This sec ion p o ides a comp ehensi e o e iew o he key heo e ical concep s u ilized in
he p esen esea ch endea o .
Sen imen analysis is a widely ecognized app oach in da a science, a ield ha in eg a es
a ious algo i hms, machine lea ning heo ies and ools o ex ac aluable insigh s om
aw da a. Cu en ly, da a science app oaches a e ex ensi ely applied in domains such as
s ock ma ke mo emen p edic ion and analysis. The mo emen o ma ke p ices is
in luenced by a my iad o online ac o s, including social media commen s, inancial news
Jou nal o Capi al
Ma ke s S udies
179
and s ock- ela ed news. Na u al language p ocessing (NLP) is employed o handle
uns uc u ed online da a, ans o ming i in o a s uc u ed o ma ha compu e s can use
in conjunc ion wi h s ock da a o p edic ma ke mo emen s (Meh a e al., 2021;Sidogi e al.,
2021;A dyan a and Sa i, 2021).
Ou p oposed model ep esen s an inno a i e app oach ha combines sen imen
analysis, adi ional ma ke indica o s like momen um and VIX and ad anced machine
lea ning echniques like long sho - e m memo y (LSTM) o o ecas mo emen s in he Ni y
50 Index. Sen imen analysis plays a pi o al ole by ex ac ing aluable insigh s om social
media o gauge in es o and public sen imen . This aspec allows us o cap u e he nuanced
emo ional esponses o ma ke pa icipan s, which o en in luence ading decisions and
ma ke ends.
In pa allel, adi ional ma ke indica o s such as momen um and he VIX p o ide
ounda ional me ics ha e lec ma ke dynamics and isk le els. These indica o s a e
c ucial in unde s anding he cu en s a e o he ma ke and i s po en ial di ec ions.
The in eg a ion o machine lea ning echniques u he enhances ou model’s
capabili ies. Machine lea ning algo i hms, known o hei abili y o lea n om da a and
adap o e ime, enable us o iden i y complex pa e ns and ela ionships wi hin as
da ase s. By inco po a ing hese adap i e capabili ies, ou model no only imp o es
p edic ion accu acy bu also adap s o e ol ing ma ke condi ions and un o eseen
e en s.
The syne gy be ween sen imen analysis, adi ional ma ke indica o s and machine
lea ning c ea es a comp ehensi e amewo k ha add esses bo h he quali a i e and
quan i a i e aspec s o ma ke o ecas ing. This holis ic app oach aims o p o ide deepe
insigh s in o ma ke beha io du ing unp eceden ed pe iods o unce ain y and ola ili y,
acili a ing mo e in o med decision-making o in es o s and s akeholde s alike.
O e all, ou model ep esen s a s ep o wa d in p edic i e analy ics o inancial ma ke s,
le e aging he powe o da a-d i en insigh s and ad anced echnology o na iga e he
complexi ies o oday’s dynamic ma ke en i onmen e ec i ely.
3.1 Sen imen analysis
Na u al language p ocessing (NLP), a b anch o a i icial in elligence, excels a
analyzing ex da a o ex ac aluable insigh s (Lin and Nuha, 2023). Wi hin his ealm,
sen imen analysis eme ges as a powe ul ool o mining public da a and unco e ing
unde lying opinions and emo ions. Le e aging NLP me hods, i classi ies he sen imen
o a ex as posi i e, nega i e o neu al (Shah e al., 2019). Among he a ailable ools,
VADER s ands ou as a ee and open-sou ce op ion ha combines dic iona y and ule-
d i en app oaches o analyze sen imen and quan i y i s in ensi y wo d by wo d. This
makes i pa icula ly adep a handling social media da a, whe e i s pola i y sco e
unc ion calcula es he emo ional leaning o each wo d (Hu o and Gilbe , 2014;Bon a
e al., 2019).
Below a e he s eps ca ied ou o sen imen analysis.
(1) To analyze he sen imen o Twi e da a, we le e age na u al language p ocessing
echniques like Vade , enabling us o de ec he emo ional one exp essed in wee s.
(2) Classi y he sen imen in o wo ypes. Type 1: public sen imen and Type 2: in es o
sen imen .
•Twee s ela ed o he gene al opic ela ing o e en s a e classi ied as public wee s,
and sen imen analysis is pe o med on hese wee s o de e mine he public
sen imen .
JCMS
8,2
180
(4) Ni y 50, bo h public and in es o sen imen (Exp-4);
(5) Ni y 50, public sen imen , in es o sen imen and VIX (Exp-5);
(6) Ni y 50, public sen imen , in es o sen imen and momen um index (Exp-6) and
(7) All ea u es (Exp-7).
This app oach allowed us o assess he inc emen al alue o each ea u e and iden i y he
combina ion ha bes cap u es he ma ke dynamics du ing hese e en s. We e alua ed
he pe o mance o each model using classi ica ion and eg ession me ics. We expec
ha inco po a ing sen imen analysis, ola ili y da a and momen um indica o s
alongside his o ical s ock p ices will lead o imp o ed p edic ion accu acy, e lec ing
he in luence o public and in es o psychology on ma ke beha io du ing pe iods o
unce ain y.
Table 1 shows esul s ob ained o all he se en di e en expe imen al ained model wi h
espec o eg ession me ics: R-squa ed, mean squa ed e o (MSE) and mean absolu e e o
(MAE), and Table 2 shows he esul s wi h espec o classi ica ion me ics: accu acy, ecall
and F1 sco e.
Geopoli ical e en s can signi ican ly impac he Indian s ock ma ke , a majo eme ging
economy highly in eg a ed wi h he global inancial sys em. This pape examines he
impac o speci ic e en s like he U i/Pulwama a acks and he Russia–Uk aine wa on he
Ni y 50 index, analyzing bo h immedia e and sus ained e ec s. While he U i/Pulwama
a acks igge ed a empo a y decline ollowed by a swi ebound (Ni y 50 down 2.2%
ollowed by a 1.7% ise) as obse ed in Figu es 3 and 4, he ongoing Russia–Uk aine wa
E en s Me ic Exp-1 Exp-2 Exp-3 Exp-4 Exp-5 Exp-6 Exp-7
E en -I R-squa ed 0.2567 0.2723 0.2763 0.3079 0.3263 0.3263 0.3482
MSE 0.0379 0.0351 0.0304 0.0268 0.0252 0.0248 0.0198
MAE 0.0968 0.0897 0.0875 0.0817 0.0810 0.0804 0.0782
E en -II R-squa ed 0.2487 0.2789 0.2772 0.3051 0.3229 0.3229 0.3471
MSE 0.0381 0.0343 0.0364 0.0282 0.0259 0.0251 0.0154
MAE 0.0951 0.0862 0.0875 0.0809 0.0804 0.0812 0.0759
E en -III R-squa ed 0.2515 0.2759 0.2763 0.3084 0.3259 0.3239 0.3391
MSE 0.0363 0.0351 0.0304 0.0289 0.0272 0.0260 0.0102
MAE 0.0942 0.0847 0.0855 0.0814 0.0815 0.0809 0.0759
Sou ce(s): Au ho s’ own wo k
E en s Me ic Exp-1 Exp-2 Exp-3 Exp-4 Exp-5 Exp-6 Exp-7
E en -I Accu acy 84.51% 87.24% 91.25% 95.01% 97.15% 96.73% 97.99%
Recall 72.18% 75.21% 78.91% 82.49% 83.71% 83.46% 87.92%
F1 Sco e 80.12% 81.27% 84.62% 89.12% 91.83% 90.04% 91.41%
E en -II Accu acy 85.01% 86.98% 91.72% 95.49% 96.72% 97.38% 98.63%
Recall 72.51% 76.12% 78.43% 82.46% 83.01% 89.85% 89.21%
F1 Sco e 80.82% 81.74% 85.01% 89.47% 90.15% 91.89% 92.35%
E en -III Accu acy 84.78% 90.21% 91.93% 95.97% 96.83% 97.62% 98.79%
Recall 72.36% 76.06% 78.31% 82.42% 83.45% 89.01% 90.12%
F1 Sco e 80.36% 85.32% 81.09% 89.73% 90.78% 91.14% 95.68%
Sou ce(s): Au ho s’ own wo k
Table 1.
Expe imen al esul s
in e ms o eg ession
me ics
Table 2.
Expe imen al esul s
in e ms o
classi ica ion me ics
Jou nal o Capi al
Ma ke s S udies
187
has caused a much mo e p olonged down u n (Ni y 50 down 815 poin s, exceeding 7%
decline wi hin he i s wo mon hs), as obse ed in Figu e 5. This dispa i y highligh s he
in luence o e en scale, economic impac and global unce ain y on ma ke beha io .
Po en ial mode a ing ac o s like go e nmen esponse and in es o con idence should
Figu e 3.
Ac ual s p edic ed
alue g aph o e en -1
Figu e 4.
Ac ual s p edic ed
alue g aph o e en -2
Figu e 5.
Ac ual s p edic ed
alue g aph o e en -3
JCMS
8,2
188
also be conside ed o a comp ehensi e unde s anding o ma ke dynamics du ing such
e en s. In eg a ing sen imen analysis alongside his da a-d i en app oach can u he
illumina e he complex in e play be ween public emo ions, news na a i es and in es o
decisions. Ul ima ely, unde s anding he nuanced ela ionship be ween geopoli ical
e en s, in es o sen imen and ma ke beha io is c ucial o in o med decision-making by
in es o s, policymake s and all s akeholde s na iga ing he e e -changing global
landscape.
5.1 Key indings
Sen imen analysis boos s accu acy: Adding public sen imen analysis o s ock da a (Exp-2)
inc eased p edic ion accu acy by 3% compa ed o jus using Ni y-50 da a (Exp-1). This
bene i g ew o 3% when in es o sen imen was included (Exp-3) and 10% when combining
bo h public and in es o sen imen (Exp-4). This highligh s he alue o sen imen analysis
in cap u ing ma ke ends.
VIX ma e : Including he VIX in Exp-5 u he imp o ed accu acy by 2% o e Exp-4.
This sugges s i s e ec i eness in gauging ma ke sen imen , as shown in Table 2.
Momen um indica o s shine oge he : In eg a ing all momen um indica o s in Exp-6
yielded an addi ional 2% accu acy gain o e Exp-4. This demons a es he syne gy o hese
indica o s in cap u ing ma ke dynamics du ing unce ain y.
Combined powe : These indings showcase he po en ial o le e aging sen imen analysis,
VIXs and momen um indica o s oge he (Exp-7). Combining hese echniques can
signi ican ly imp o e s ock p edic ion accu acy by 14% compa ed o jus using Ni y-50
da a (Exp-1).
5.2 Implica ions o indings
5.2.1 Sen imen analysis cap u es in es o psychology (RQ1). The s udy success ully
add essed he ole o sen imen analysis in iden i ying in es o beha io du ing geopoli ical
ensions (RQ1). Public sen imen analysis alone imp o ed p edic ion accu acy by 3% (Exp-2
s Exp-1), demons a ing i s abili y o cap u e b oad ma ke ends. No ably, including
in es o sen imen analysis yielded an addi ional boos o 3% (Exp-3 s Exp-1). This inding
di ec ly suppo s he objec i e o de eloping a hyb id model ha le e ages social media
analysis o unde s and in es o con idence o ne ousness du ing such e en s. The g ea es
imp o emen (10%, Exp-4 s Exp-1) came om combining bo h public and in es o
sen imen , highligh ing he syne gy achie ed by inco po a ing hese complemen a y
sou ces.
5.2.2 Ma ke indica o s enhance model pe o mance (RQ2). The esea ch also
in es iga ed he e ec i eness o adi ional ma ke indica o s in cap u ing he ma ke ’s
esponse o geopoli ical ensions (RQ2). The VIX p o ed aluable, inc easing p edic ion
accu acy by 2% (Exp-5 s Exp-4). This sugges s ha inco po a ing ola ili y da a a e
e ec i e in gauging ma ke sen imen du ing pe iods o unce ain y. Addi ionally,
in eg a ing all momen um indica o s (MACD, RSI, e c.) esul ed in a u he 2% accu acy
gain (Exp-6 s Exp-4). This inding aligns wi h he objec i e o using ma ke indica o s o
unde s and ma ke dynamics du ing hese e en s.
5.2.3 Hyb id model deli e s enhanced p edic ion accu acy (RQ3). The co e ou come o he
esea ch di ec ly add esses RQ3. The signi ican 14% imp o emen in p edic ion
accu acy achie ed by he combined model (Exp-7) compa ed o he baseline model using
only Ni y-50 da a (Exp-1) p o ides a clea answe . This inding demons a es ha he
hyb id model, inco po a ing sen imen analysis and ma ke indica o s, is demons ably
mo e accu a e in p edic ing sho - e m s ock ma ke mo emen s du ing geopoli ical
ensions.
Jou nal o Capi al
Ma ke s S udies
189
By quan i ying he imp o emen (14%), he esea ch di ec ly add esses he ques ion o
how accu a ely he model p edic s hese mo emen s. This ou come alida es he objec i e o
e alua ing he model’s e ec i eness and suppo s he hypo hesis ha he hyb id app oach
ou pe o ms exis ing models in such si ua ions.
5.3 Compa ison wi h exis ing wo k
We ha e compa ed ou p edic ion esul s wi h he ollowing wo o he la es published
ela ed wo ks.
Wang e al. (2020) discussed ega ding he in es o sen imen and s ock p ice mo emen .
The au ho s ha e used Eas Money, he mos popula online o um o China o S ocks, o
analyzing he in es o sen imen , on he s ock mo emen s o CSI300 Index which is aded in
majo s ock exchanges o China. The au ho s ha e analyzed he e ec o in es o sen imen
on ading olume, s ock p ice, o de imbalance o big ade e c. We ha e conside ed his o
compa ison as he au ho s used he sen imen analysis o online in es o sen imen wi h
LSTM is used as sen imen classi ie .
Das e al. (2022) conside ed ou di e en news sou ces like Facebook commen s, inancial
news and s ock- ela ed a icles om Economic Times and Twi e da a and se en di e en
echniques like VADER, Logis ic Reg ession, Lough an–McDonald, Hen y, Tex Blob,
Linea SVC and S an o d a e used o sen imen analysis. Sen imen analysis was pe o med
on indi idual da a sou ce as well as he combina ion o he di e en da a sou ces. We ha e
conside ed his o compa ison as he au ho s used mul iple sen imen analysis echniques
ac oss mul iple da a sou ces o p edic ion.
Table 3 shows he compa ison o his p oposed esea ch wo k wi h wo o he la es
published ela ed wo ks. F om he compa ison wi h he ela ed wo ks, i is e iden ha he
p oposed wo k achie es highe accu acy in p edic ion as i uses mul iple pa ame e s like
sen imen analysis, momen um indica o and VIX da a along wi h he s ock da a.
Wo k
done Da a sou ces
S ock
da a
Tool o sen imen
analysis
P edic ion
me hod Accu acy
Wang e al Online Fo um CSI
300
LSTM Machine
Lea ning
Algo i hms
A g: 77.83%
Das e al S ock- ela ed a icles
headlines om
“Economic Times,”
Twee s om Twi e ,
Financial news om
“Economic Times”
and Facebook
commen s
Ni y
50
VADER, Logis ic
Reg ession,
Lough an–McDonald,
Hen y, Tex Blob,
Linea SVC and
S an o d
LSTM Linea SVC:
98.32%
Logis ic
Reg ession:
97.67%
VADER:
96.85%
Lough an–
McDonald:
94.78% Hen y:
96.36%
Tex Blob:
96.48%
S an o d:
96.57%
P oposed
wo k
Twi e , VIX and
Momen um
Ni y
50
VADER CNN-
BDLSTM
A g: 98.47%
Sou ce(s): Au ho s’ own wo k
Table 3.
Compa ison wi h
exis ing wo k
JCMS
8,2
190
6. Conclusion and p ac ical implica ions
In conclusion, his pape has no only con i med he signi ican impac o geopoli ical
ensions on he Indian s ock ma ke bu also demons a ed he c ucial ole o sen imen
analysis in unde s anding in es o beha io du ing hese e en s. By analyzing social media
sen imen da a, we de eloped a no el p edic ion model ha inco po a es sen imen alongside
ola ili y and momen um indica o s o imp o ed accu acy.
We success ully de eloped a p edic i e model ha u ilizes sen imen analysis, VIX and
momen um indica o o o ecas s ock ma ke alues du ing geopoli ical ensions wi h an
accu acy o 98.47%. Ou esea ch shows ha including mo e han jus his o ical Ni y 50
da a signi ican ly imp o es ou model’s abili y o p edic s ock mo emen s du ing
geopoli ical ensions. By inco po a ing public and in es o sen imen analysis wi h s ock
da a (Exp-4), we achie ed a 10% accu acy boos compa ed o using only Ni y 50 da a (Exp-
1). Adding he VIX and momen um indica o s (Exp-7) u he imp o ed accu acy by 4%.
This highligh s he aluable ole hese ac o s play in cap u ing ma ke sen imen and
dynamics du ing unce ain imes.
6.1 Limi a ions
While ou cu en esea ch demons a es subs an ial accu acy o p edic ion o he s ock
ma ke du ing geopoli ical e en s, u u e i e a ions will explo e expanding he model’s
capabili ies:
(1) Da a di e si ica ion: In eg a ing addi ional da a sou ces like news a icles, o icial
epo s and b oade social media pla o ms can o e a mo e comp ehensi e iew o
public and expe sen imen , po en ially enhancing o ecas ing accu acy.
(2) Technical indica o expansion: Including me ics like ma ke b ead h and o eign
in es o ac i i y can p o ide deepe insigh s in o ma ke in e nals and in es o
beha io , imp o ing he model’s abili y o cap u e he nuanced dynamics o
geopoli ical ension-d i en ma ke luc ua ions.
(3) Add essing limi a ions: Op imizing indica o se ings and using hese indica o s in
conjunc ion wi h o he echnical and undamen al analysis ools can p o ide a
mo e comp ehensi e iew o he ma ke and educe he isk o elying on ew
indica o s.
6.2 Recommenda ions
6.2.1 Fo esea che s.
(1) In eg a e sen imen analysis da a in o o ecas ing models o gain deepe insigh s in o
socie al in luences on economic ac i i y and
(2) De elop mo e accu a e o ecas ing models leading o be e p edic ions o u u e
economic ends.
6.2.2 Fo in es o s.
(1) U ilize sen imen analysis ools o gain a b oade unde s anding o public in e es and
make mo e in o med in es men decisions, especially du ing pe iods o high ma ke
ola ili y.
6.2.3 Fo policymake s.
(1) Le e age models in o med by sen imen analysis da a o unde s and socie al
conce ns mo e e ec i ely and
Jou nal o Capi al
Ma ke s S udies
191
(2) C a be e esponse s a egies and measu es o s abilize inancial ma ke s du ing
imes o un es .
6.2.4 In es o bene i s.
(1) P ecise s ock selec ion and isk educ ion: This model o e s a aluable edge o
in es o s, especially du ing ma ke u bulence. By combining sen imen analysis
wi h momen um indica o s, i p o ides a comp ehensi e iew o po en ial s ock
mo emen s. This allows o mo e accu a e s ock selec ion, ea lie iden i ica ion o
ends and, ul ima ely, educed in es men isk.
(2) Fu u e-p oo ing he model: The model’s success pa es he way o exci ing
ad ancemen s. This includes in eg a ion in o au oma ed ading s a egies and
a ac ing pa ne ships o in es men s o expand i s applica ion. These de elopmen s
hold he po en ial o signi ican ly imp o e s ock p edic ion accu acy and ma ke
sen imen insigh s. Ul ima ely, his ansla es o be e in es men decisions and
po en ially highe e u ns o in es o s.
Re e ences
Ahe n, K.R. and Sosyu a, D. (2015), “Rumo has i : sensa ionalism in inancial media”, Re iew o
Financial S udies, Vol. 28 No. 7, pp. 2050-2093, doi: 10.1093/ s/hh 006.
Alqah ani, A., Bou i, E. and Vo, X.V. (2020), “P edic abili y o GCC s ock e u ns: he ole o
geopoli ical isk and c ude oil e u ns”, Economic Analysis and Policy, Vol. 68, pp. 239-249,
doi: 10.1016/j.eap.2020.09.017.
A dyan a, E.I. and Sa i, H. (2021), “A p edic ion o s ock p ice mo emen s using suppo ec o
machines in Indonesia”, The Jou nal o Asian Finance, Economics and Business, Vol. 8 No. 8,
pp. 399-407, doi: 10.13106/JAFEB.2021.VOL8.NO8.0399.
Bake , M. and Wu gle , J. (2000), “The p ice o ea : es ing he limi s o a ional ma ke s”, Jou nal o
Poli ical Economy, Vol. 08 No. 3, pp. 670-711.
Bane jee, A.V. (1992), “A simple model o he d beha io ”, Qua e ly Jou nal o Economics, Vol. 107
No. 3, pp. 797-817, doi: 10.2307/2118364.
Ba be is, N.C. (2013), “Thi y yea s o p ospec heo y in economics: a e iew and assessmen ”, The
Jou nal o Economic Pe spec i es, Vol. 27 No. 1, pp. 173-196, doi: 10.1257/jep.27.1.173.
Bekae , G., Hoe o a, V. and Muellbaue , M. (2014), “Measu ing in e na ional poli ical isk: why, how,
and o whom?”, Jou nal o Empi ical Finance, Vol. 31, pp. 8-32.
Be k, J. and DeMa zo, P. (2017), Co po a e Finance, Global Edi ion.
Bollen, J., Mao, H. and Zeng, X. (2011), “Twi e mood p edic s he s ock ma ke ”, Jou nal o
Compu a ional Science, Vol. 2 No. 1, pp. 1-8, doi: 10.1016/j.jocs.2010.12.007.
Bon a, V., Kuma esh, N. and Jana dhan, N. (2019), “A comp ehensi e s udy on lexicon based
app oaches o sen imen analysis”, Asian Jou nal o Compu e Science Technology, Vol. 8
No. 2, pp. 1-6, doi: 10.51983/ajcs -2019.8.s2.2037.
Cookson, J.A., Engelbe g, J. and Mullins, W. (2020), “Echo chambe s: social media and poli ical
ex emism”, Managemen Science, Vol. 66 No. 11, pp. 4945-4964, doi: 10.1287/mnsc.
2019.3481.
Das, S. and Vasileios, A. (2014), “News sen imen and s ock ma ke ola ili y: an empi ical analysis o
21 coun ies”, Jou nal o Banking and Finance, Vol. 38 No. 3, pp. 505-520.
Das, N., Sadhukhan, B., Cha e jee, T. and Chak aba i, S. (2022), “E ec o public sen imen on s ock
ma ke mo emen p edic ion du ing he COVID-19 ou b eak”, Social ne wo k analysis and
mining, Vol. 12 No. 1, p. 92, doi: 10.1007/s13278-022-00919-3.
JCMS
8,2
192
Fama, E.F. (1970), “E icien capi al ma ke s: a e iew o heo y and empi ical wo k”, The Jou nal o
Finance, Vol. 25 No. 2, pp. 383-417, doi: 10.2307/2325486.
Fama, E.F. (1991), “E icien capi al ma ke s: II”, The Jou nal o Finance, Vol. 46 No. 5, pp. 1575-1617,
doi: 10.1111/j.1540-6261.1991. b04636.x.
Fio illo, P., Meles, A., Pelleg ino, L.R. and Ve doli a, V. (2023), “Geopoli ical isk and s ock liquidi y”,
Finance Resea ch Le e s, Vol. 54, 103687, doi: 10.1016/j. l.2023.103687.
Hachicha, F. (2023), “Sen imen in es o , exchange a es, geopoli ical isk and de eloping s ock
ma ke : e idence o co-mo emen s in he ime- equency domain du ing RussiaUk aine wa ”,
Re iew o Beha io al Finance, Vol. 16 No. 3, pp. 486-509, doi: 10.1108/ b -04-2023-0119.
Heds €
om, A., Zelande , N., Jun ila, J. and Uddin, G.S. (2020), “Eme ging ma ke con agion unde
geopoli ical unce ain y”, Eme ging Ma ke s Finance and T ade, Vol. 56 No. 6, pp. 1377-1401,
doi: 10.1080/1540496x.2018.1562895.
Hi shlei e , D. and Teoh, S.H. (2003), “He d beha io and cascading in capi al ma ke s: a e iew and
syn hesis”, Eu opean Financial Managemen , Vol. 9 No. 1, pp. 25-66, doi: 10.1111/1468-036X.00207.
Hu o, C. and Gilbe , E. (2014), “VADER: a pa simonious ule-based model o sen imen analysis o
social media ex ”, P oceedings o he In e na ional AAAI Con e ence on Web and Social Media,
Vol. 8, pp. 216-225, doi: 10.1609/icwsm. 8i1.14550.
Kahneman, D. and T e sky, A. (1979), “P ospec heo y: an analysis o decision unde isk”,
Econome ica, Vol. 47 No. 2, pp. 263-291, doi: 10.2307/1914185.
Lee, C.M. and Wu, C. (2016), “In es o sen imen and s ock p ice synch onici y: e idence om he US
ma ke ”, Jou nal o Financial Economics, Vol. 120 No. 1, pp. 1-30.
Li, T., Ruan, W. and Zhang, Y. (2019), “The impac s o social media on he inancial pe o mance o ini ial
public o e ings”, Jou nal o Financial S abili y, Vol. 42, pp. 125-138, doi: 10.1016/j.j s.2019.04.005.
Lin, C.H. and Nuha, U. (2023), “Sen imen analysis o Indonesian da ase s based on a hyb id deep-
lea ning s a egy”, Jou nal o Big Da a, Vol. 10 No. 1, p. 88, doi: 10.1186/s40537-023-00782-9.
Lough an, T. and McDonald, B. (2011), “When is a liabili y no a liabili y? Tex ual analysis,
dic iona ies, and 10-Ks”, The Jou nal o Finance, Vol. 66 No. 1, pp. 35-65, doi: 10.1111/j.1540-
6261.2010.01625.x.
Manu ung, A.H., Budiha o, W. and P abowo, H. (2018), “Algo i hm and modeling o s ock p ices
o ecas ing based on long sho - e m memo y (LSTM)”, ICIC Exp ess Le e s, Vol. 12 No. 12,
pp. 1277-1283.
Meh a, P., Pandya, S. and Ko echa, K. (2021), “Ha es ing social media sen imen analysis o enhance
s ock ma ke p edic ion using deep lea ning”, Pee J Compu e Science, Vol. 7, e476, doi: 10.
7717/pee j-cs.476.
Plakanda as, V., Gogas, P. and Papadimi iou, T. (2018), “The e ec s o geopoli ical unce ain y in
o ecas ing inancial ma ke s: a machine lea ning app oach”, Algo i hms, Vol. 12 No. 1, p. 1,
doi: 10.3390/a12010001.
Salisu, A.A., Lasisi, L. and Tchankam, J.P. (2022a), “His o ical geopoli ical isk and he beha iou o
s ock e u ns in ad anced economies”, The Eu opean Jou nal o Finance, Vol. 28 No. 9,
pp. 889-906, doi: 10.1080/1351847x.2021.1968467.
Salisu, A.A., Ogbonna, A.E., Lasisi, L. and Olani an, A. (2022b), “Geopoli ical isk and s ock ma ke
ola ili y in eme ging ma ke s: a GARCH–MIDAS app oach”, The No h Ame ican Jou nal o
Economics and Finance, Vol. 62, 101755, doi: 10.1016/j.naje .2022.101755.
Segnon, M., Gup a, R. and Wil ling, B. (2023), “Fo ecas ing s ock ma ke ola ili y wi h egime-
swi ching GARCH-MIDAS: he ole o geopoli ical isks”, In e na ional Jou nal o Fo ecas ing,
Vol. 40 No. 1, pp. 29-43, doi: 10.1016/j.ij o ecas .2022.11.007.
Shah, D., Isah, H. and Zulke nine, F. (2019), “S ock ma ke analysis: a e iew and axonomy o
p edic ion echniques”, In e na ional Jou nal o Financial S udies, Vol. 7 No. 2, p. 26, doi: 10.
3390/ij s7020026.
Jou nal o Capi al
Ma ke s S udies
193
Shehzad, S. and Malik, F. (2019), “In es o sen imen and s ock ma ke ola ili y: a no el app oach
using machine lea ning”, The Jou nal o Beha io al Finance, Vol. 20 No. 1, pp. 1-19.
Shille , R.J. (2000), I a ional Exube ance, P ince on Uni e si y P ess, New Yo k.
Shlei e , A. and Summe s, L.H. (1990), “The noise ade app oach o inance”, The Jou nal o
Economic Pe spec i es, Vol. 4 No. 2, pp. 19-33, doi: 10.1257/jep.4.2.19.
Sidogi, T., Mbu ha, R. and Ma wala, T. (2021), “S ock p ice p edic ion using sen imen analysis”,
IEEE In e na ional Con e ence on Sys ems, Man, and Cybe ne ics (SMC), Melbou ne, pp. 46-51,
doi: 10.1109/SMC52423.2021.9659283.
Te lock, P.C. (2007), “Gi ing con en o in es o sen imen : he ole o media in he s ock ma ke ”,
The Jou nal o Finance, Vol. 62 No. 3, pp. 1139-1168, doi: 10.1111/j.1540-6261.2007.01232.x.
T iki, M.B. and Maa oug, A.B. (2021), “The GOLD ma ke as a sa e ha en agains he s ock ma ke
unce ain y: e idence om geopoli ical isk”, Resou ces Policy, Vol. 70, 101872, doi: 10.1016/j.
esou pol.2020.101872.
Wang, G., Yu, G. and Shen, X. (2020), “The e ec o online in es o sen imen on s ock mo emen s:
an ls m app oach”, Complexi y, pp. 1-11, doi: 10.1155/2020/4754025.
Wa s, D.J. (2002), “A simple model o global cascades on andom ne wo ks”, P oceedings o he
Na ional Academy o Sciences, Vol. 99, pp. 5766-5771, doi: 10.1073/pnas.082090499.
Co esponding au ho
S i a sa Maddodi can be con ac ed a : [email p o ec ed]
Fo ins uc ions on how o o de ep in s o his a icle, please isi ou websi e:
www.eme aldg ouppublishing.com/licensing/ ep in s.h m
O con ac us o u he de ails: [email p o ec ed]
JCMS
8,2
194
