Predicting the economic impact of the COVID-19 pandemic in the United Kingdom using time-series mining

Rakha, Ahmed e al. A icle P edic ing he economic impac o he COVID-19 pandemic in he Uni ed Kingdom using ime-se ies mining Economies P o ided in Coope a ion wi h: MDPI – Mul idisciplina y Digi al Publishing Ins i u e, Basel Sugges ed Ci a ion: Rakha, Ahmed e al. (2021) : P edic ing he economic impac o he COVID-19 pandemic in he Uni ed Kingdom using ime-se ies mining, Economies, ISSN 2227-7099, MDPI, Basel, Vol. 9, Iss. 4, pp. 1-19, h ps://doi.o g/10.3390/economies9040137 This Ve sion is a ailable a : h ps://hdl.handle.ne /10419/257295 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/ economies A icle P edic ing he Economic Impac o he COVID-19 Pandemic in he Uni ed Kingdom Using Time-Se ies Mining Ahmed Rakha 1,†, Hansi He ia achchi 2,†, Dina Rady 3, Mohamed Medha Gabe 2,4, Emad Rakha 5 and Mohammed M. Abdelsamea 2,6,*   Ci a ion: Rakha, Ahmed, Hansi He ia achchi, Dina Rady, Mohamed Medha Gabe , Emad Rakha, and Mohammed M. Abdelsamea. 2021. P edic ing he Economic Impac o he COVID-19 Pandemic in he Uni ed Kingdom Using Time-Se ies Mining. Economies 9: 137. h ps:// doi.o g/10.3390/economies9040137 Academic Edi o s: Edwa d C. Hoang and Joydeep Bha acha ya Recei ed: 29 July 2021 Accep ed: 22 Sep embe 2021 Published: 27 Sep embe 2021 Publishe ’s No e: MDPI s ays neu al wi h ega d o ju isdic ional claims in published maps and ins i u ional a il- ia ions. Copy igh : © 2021 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/). 1School o Business, No ingham Uni e si y, No ingham NG8 1BB, UK; [email p o ec ed] 2School o Compu ing and Digi al Technology, Bi mingham Ci y Uni e si y, Bi mingham B4 7XG, UK; [email p o ec ed] (H.H.); [email p o ec ed] (M.M.G.) 3Economics Depa men , Geo ge Washing on Uni e si y, Washing on, DC 20052, USA; [email p o ec ed] 4Facul y o Compu e Science and Enginee ing, Galala Uni e si y, Suez 435611, Egyp 5School o Medicine, No ingham Uni e si y, No ingham NG8 1BB, UK; [email p o ec ed] 6Facul y o Compu e s and In o ma ion, Assiu Uni e si y, Assiu 71515, Egyp *Co espondence: [email p o ec ed] † Deno es equal con ibu ion. Abs ac : The COVID-19 pandemic has b ough economic ac i i y o a nea s ands ill as many coun- ies imposed e y s ic es ic ions on mo emen o hal he sp ead o he i us. This s udy aims a assessing he economic impac s o COVID-19 in he Uni ed Kingdom (UK) using a i icial in elligence (AI) and da a om p e ious economic c ises o p edic u u e economic impac s. The mac oeconomic indica o s, g oss domes ic p oduc s (GDP) and GDP g ow h, and da a on he pe o mance o h ee p ima y indus ies in he UK ( he cons uc ion, p oduc ion and se ice indus ies) we e analysed using a compa ison wi h he pa e n o p e ious economic c ises. In his esea ch, we expe imen ed wi h he e ec i eness o bo h con inuous and ca ego ical ime-se ies o ecas ing on p edic ing u u e alues o gene a e mo e accu a e and use ul esul s in he economic domain. Con inuous alue p edic ions indica e ha GDP g ow h in 2021 will emain s eady, bu a a ound − 8.5% con ac ion, compa ed o he baseline igu es be o e he pandemic. Fu he , he ca ego ical p edic ions indica e ha he e will be no qua e ly d op in GDP ollowing he i s qua e o 2021. This s udy p o ided e idence-based da a on he economic e ec s o COVID-19 ha can be used o plan necessa y eco e y p ocedu es and o ake app op ia e ac ions o suppo he economy. Keywo ds: COVID-19; economic impac s; UK; indus y; g oss domes ic p oduc s 1. In oduc ion The COVID-19 pandemic has igge ed a massi e heal h c isis ac oss he globe and o ced many coun ies o ake se e e p e en i e measu es o mi iga e he sp ead o he i us including en o cing na ional lockdowns and social dis ancing measu es ( Huang e al. 2020 ). Subsequen ly, many businesses we e pushed o he b ink o collapse and he na ional economy su e ed signi ican ly in almos all aspec s (Song and Zhou 2020). Al hough all coun ies and businesses a e a ec ed, he magni ude o he impac s ha e a ied massi ely. The du a ion and he pa e n o he lockdown and he social dis ance measu es imposed by he go e nmen s o limi he sp ead o he pandemic has caused asymme ic e ec s no only be ween coun ies bu also be ween business sec o s and in he demand-supply chains wi hin each coun y. This asymme y o he economic shock may indica e ha p edic ing he g oss domes ic p oduc s (GDP) will be challenging and could be unexpec edly di e en om he o ecas ed igu es (BFPG 2020). The i s Uni ed Kingdom (UK) go e nmen ad ice on social dis ancing was published on 12 Ma ch 2020, be o e a o mal “lockdown”, which was announced on 23 Ma ch 2020. As in o he a ec ed coun ies, his led o a all in he consume demand and business Economies 2021,9, 137. h ps://doi.o g/10.3390/economies9040137 h ps://www.mdpi.com/jou nal/economies Economies 2021,9, 137 2 o 19 and ac o y closu es, as well as supply chain dis up ions. Al hough ea ly da a indica ed ha he UK is less a ec ed compa ed o some o he Eu opean coun ies, as he pandemic de eloped he pe o mance o he UK ela i e o hese o he Eu opean coun ies wo sened (Boissay and Rungcha oenki kul 2020). While ini ially much emphasis had been placed upon he UK’s a ou able pe o mance compa ed o o he coun ies, wi h ime he UK go e nmen has a emp ed o p o ide a posi i e ep esen a ion o ends o he disease sp ead (Balm o d e al. 2020;Paisley 2020). The uns able and changing si ua ions make o ecas ing he economic pe o mance and p edic ing he immedia e and long- e m economic impac s challenging o many easons. These include he global na u e o he pandemic wi h dis up ion o he demand and supply chains, anspo a ion es ic ions, he na ional measu es aken, which a ec ing almos all business, and ex ended du a ion o he pandemic, he huge job losses, and he signi ican ise in he unemploymen a es, he way educa ion and o he business a e deli e ed and he lack o simila models ha can be used o p edic he u u e pe o mance (Sake e al. 2004). The s agna ion o business, he d op in he na ional ax income and GDP will limi he abili y o he go e nmen o eac and ake emedial ac ions o ensu e a balanced eco e y o all sec o s (Child 2021). The e o e, da a on he pe o mance o di e en sec o s, and accu a e o ecas ing o he pe o mance and eco e y a e needed o allow u u e planning and dis ibu ion o he al eady s e ched esou ces (Kim e al. 2021). In addi ion, a obus o ecas ing model ha can conside he mul idimensional na u e o he pandemic, and lea n om p e ious economic c ises and pandemics is needed. A i icial in elligence (AI) has been success ul in a a ie y o ields including com- pu e ision, obo ics, aud de ec ion, d ug disco e y, and epidemiology (Ahuja 2019; Ib ahim e al. 2020 ;Wahl e al. 2018;A slan and Benke 2021;Fu e al. 2019). The e is a g ea hope ha AI app oaches can be a key o suppo ing s udies o COVID-19 and u u e c ises o build dynamic and esilien o ecas ing models ha can u ilise he da a o p e ious pandemics and o he economic c ises. De eloping in elligen sys ems ha can help go - e nmen s p edic ing pe o mance and economic ou comes, o e ing solu ion models can be e y help ul in such mul idimensional c ises and help o ackle u u e c ises and challenges. This s udy aims o cha ac e ise he impac s o COVID-19 on he UK mac oeconomic indica o s wi h compa ison o p e ious economic c ises in he ecen his o y and o u ilise AI o p edic he u u e economic impac s o COVID-19 in he UK. We mainly ocused on bo h con inuous and ca ego ical ime se ies o ecas ing me hods o p edic u u e measu es. E en hough he e was a high endency o use con inuous o ecas ing me hods o such p edic ions, we in ol ed ca ego ical o ecas ing o mi iga e he impac o unexpec ed a ia ions due o he pandemic si ua ion on p edic ions. Da a on he mac oeconomic indica o s, g oss domes ic p oduc s (GDP) and GDP g ow h, and da a on he pe o mance o h ee p ima y indus ies in he UK (cons uc ion, p oduc ion, and se ice) we e analysed. 2. Li e a u e Re iew In his sec ion, we p o ide an o e iew o some o he apidly g owing li e a u e on he economic impac o COVID-19 o syn hesize he insigh s eme ging om hese s udies, allow o a compa a i e e iew, and ela e he esul s o ou s udy. - The impac o COVID-19, associa ed beha iou s and policies on he UK economy has been s udied by a compu able gene al equilib ium model (Keogh-B own e al. 2020). The au ho s o he s udy used a compu able gene al equilib ium (CGE) model linked o a popula ion-wide epidemiological demog aphic model o analyse he po en ial mac oeconomic impac o he COVID-19 on he UK economy. The s udy ound ou ha he mi iga ion s a egies ha he go e nmen imposes o 12 weeks would educe case a ali ies by 29% bu impose a o al cos o he economy o 13.5% o GDP, due o business closu e and labou loss om wo king pa en s du ing school closu es. The mi iga ion s a egies o ace he pandemic ha las s o a longe pe iod would educe dea hs by 95% bu would inc ease he o al cos o he UK economy o 29.2% o GDP, whe e 7.3% o GDP is due o school closu es and 21.9% o GDP o business closu es. Economies 2021,9, 137 3 o 19 The au ho s concluded ha COVID-19 has he po en ial o impose unp eceden ed economic impac on he UK economy and hose impac s a e likely o be domina ed by he indi ec cos s o mi iga ion o he pandemic. The du a ion o he policies o mi iga ing he sp ead o he disease is a key o de e mining he economic impac . - The impac s o COVID-19 and B exi on he UK economy we e in es iga ed (De Lyon and Dhing a 2021) using eal- ime da a p o ided by he Con ede a ion o B i ish Indus y (CBI) o assess he impac o COVID-19 on i ms’ ac i i ies and p edic ed he economic consequences o he p e en i e measu es ha he go e nmen imposed. The au ho s ha e demons a ed ha he pandemic has caused a sha p educ ion in he g ow h a e o nominal wages. Be o e he pandemic, in Janua y 2020 nominal wages had g own by 3.1% o e he p e ious 12-mon h pe iod compa ed wi h 0.9% a yea la e . Howe e , a e age p ices ha e con inued o inc ease 0.5% h oughou he pandemic, wi h a p oduce p ice in la ion o 0.6% epo ed by he ONS o all manu ac u ed p oduc s. Lowe wages o en educe demand o goods and se ices and pu downwa d p essu e on p ices. Ye despi e lowe ea nings g ow h since he pandemic, p ices ha e inc eased. This is pa ially explained by ising a e age cos s, a leas in he manu ac u ing sec o . The inc ease in cos s was mainly due o he inc eased inpu cos s by inc easing ba ie s o ade be ween he UK and he EU imposed by B exi in addi ion o delays a he bo de , bu densome adminis a i e cos s, he adop ed new echnologies and managemen p ac ices due o he pandemic. These cos s con ibu ed o he sha p all in UK ade in 2021, leading o ising cos s, highe p ices, and educed compe i i eness. - Ano he s udy (S ephens e al. 2020) has been conduc ed on he UK economy by analysing he o e all impac s o he (COVID-19) pandemic on GDP du ing July 2020 using an online ques ionnai e, and he Mon hly Business Su ey (MBS) as he p i- ma y da a sou ce o 75% o p oduc ion indus ies and 50% o se ices indus ies. The au ho s concluded ha o al se ices ou pu du ing July 2020 we e signi ican ly a ec ed by he COVID-19 pandemic and ell a 12.6% below he Feb ua y 2020 le el, he las ull mon h o “no mal” ope a ing condi ions. The anspo sec o emained one o he bigges a eas o se ices a ec ed by he co ona i us as people ha e been homewo king, land anspo has been a ec ed by a educ ion in commu e s, espe- cially he London Unde g ound, which in July 2020 was a 23% o i s usage in July 2019. P oduc ion ou pu du ing July 2020 was a 7.0% below he le el o Feb ua y 2020, he las ull mon h o “no mal” ope a ing condi ions. In July 2020, wea ing appa el was 32.9% below he Feb ua y 2020 le el. Cons uc ion ou pu was 11.6% below i s Feb ua y 2020 le el.A i e-pa amewo k (O’Donnell and Begg 2020) has been p oposed o assessing why he UK pe o med poo ly compa ed wi h o he coun ies ega ding he medical, social and economic challenges b ough abou by he COVID-19. They sugges ed ha he e ha e been p oblems as o how e idence abou he pandemic and i s an icipa ed e ec s has been collec ed, p ocessed and ci cula ed. Subsequen ly, his a ec ed policy make ’s abili y o e alua e he isk and p o iding hem wi h inadequa e in o ma ion o ace he mul idimensional challenges b ough abou by he pandemic wi h he p ope policy esponse. The s udy also e ealed ha he UK’s ins i u ional se ing may ha e been a d i e o his poo pe o mance owa ds he pandemic. The au ho s a gued ha UK policymake s may ha e o e - elied on he medical sciences a he expense o o he (social) scien i ic e idence. Ano he s udy emphasised he impo ance o widening o ganisa ional decision making as an ap- p oach (Child 2021) o add ess he p oblems ha we e b ough abou by he pandemic. The au ho de ined ‘o ganisa ional pa icipa ion’ and p o ided a amewo k o he di e en o ms o pa icipa ion, concluding ha a combina ion o co-de e mina ion and wo kplace sel -managemen is likely o ha e he mos consequen ial e ec s. The au ho a gued ha he public esponse o he c isis migh lead o a cons uc i e way o wa d o deal wi h he pandemic and emphasised he need o de elop e ec i e sys ems o pa icipa ion a all le els o human o ganiza ion. Economies 2021,9, 137 4 o 19 - Mo eo e , he main challenges acing policymake s, when ying o balance be ween inding jobs o he layou wo ke s and ge ing hem back o he igh jobs, we e highligh ed in he s udy by (Cos a Dias e al. 2020). These au ho s concluded ha he na u e o he economic shock associa ed wi h he COVID-19 pandemic is highly unusual; i has led o no only a sha p all in labou demand in many sec o s o he economy, much mo e han in a ypical down u n o slowdown in economic ac i i y, bu also a adical change in he ypes o economic ac i i y in he UK which will cause employe s o expec la ge changes o hei wo k o ce o e he nex yea . This would lead he go e nmen o ake a o wa d-looking app oach and equip people wi h he skills ha a e likely o be in demand in he u u e, p o ided a po en ial shi owa ds e-comme ce, and he need o mo e o a ne -ze o, which is he a ge o he UK o each by 2050. - The Impac o COVID-19 on sha e p ices in he UK (G i i h e al. 2020) was in es i- ga ed by Rachel G i i h and colleagues. These au ho s desc ibed how he impac s o he COVID-19 pandemic ha e a ied ac oss indus ies, using da a on he sha e p ices o he i ms lis ed on he London S ock Exchange. The alue o aded sha es in he s ock ma ke e lec s no only how well a company is doing oday, bu also how well i is expec ed o do in he u u e. In addi ion, sha e p ices e lec ma ke expec a ions abou changes in inal demand, in e media e demand, and es ic ions in supply. The au ho s concluded ha he indus ies ha ha e been hi he ha des include ou ism and leisu e, ossil uels p oduc ion and dis ibu ion, banking, insu ance, and e aile s. On he o he hand, o he indus ies ha e ou pe o med he ma ke , including ood and d ug manu ac u e s and e aile s, u ili ies, high- ech manu ac u ing, obacco, and i ms in medical and bio ech esea ch. As social dis ancing measu es con inue, capi al-in ensi e i ms migh lose he skills and expe ience o hei wo ke s. I hose i ms a e no able o educe hei cos s, hey a e mo e likely o su e in he u u e i no backed by go e nmen suppo . F om he abo e sec ion, we conclude ha mos o he esea ch ha in es iga ed he impac o he COVID-19 on he UK economy has ocused on he di e en sec o s o he economy and he o e all GDP g ow h a e. Howe e , none o hese s udies used AI model o o ecas he pa e n o change in he u u e and o compa e wi h he pa e n o p e ious economic c ises, which highligh s he impo ance and uniqueness o ou s udy. 3. Me hodology 3.1. Sample Sec ion The da a o his s udy we e collec ed om he O ice o Na ional S a is ics (ONS 2021) , which is he UK’s la ges independen p oduce o da a. I was essen ial o conside he pandemic will a ec a ious sec o s di e en ly, and each indus y will impac he econ- omy di e en ly (Coakley e al. 2014). In his s udy, da a on h ee indus ies ( he se ice, manu ac u ing and cons uc ion indus ies) and g oss domes ic p oduc (GDP) we e col- lec ed on a mon hly, qua e ly, and yea ly basis. P edic ion analysis u ilised he qua e ly da a collec ed om Janua y 1997 o Janua y 2021. Da a on GDP we e also collec ed om 1947 o Feb ua y 2021 o allow he compa ison wi h mo e p e ious ou b eaks and o he calcula ion o economic g ow h (ONS 2021). Fo he se ice indus y, he da a we e col- lec ed om he Mon hly Business Su ey (MBS) epo o u no e o se ices indus ies. The o al ou pu o all o he sec o s and 37 di e en sec ions inside he indus y we e obse ed (ONS 2021) . The da a o he manu ac u ing indus y we e collec ed om he MBS u no e o he p oduc ion indus ies. The o al u no e o he p oduc ion and man- u ac u ing sec o s and 45 di e en sec o s inside he indus ies we e obse ed (ONS 2021) . Fo he cons uc ion indus y, da a o he ou pu s, seasonally and non-seasonally adjus ed we e obse ed. Fu he mo e, yea ly and qua e ly g ow h was obse ed (ONS 2021) . Fo his s udy, a nowcas ing model based on a dynamic ac o model was used o p edic qua e ly GDP o 2021 by analysing mon hly GDP da a. This model uses eal- ime and Economies 2021,9, 137 5 o 19 high- equency da a o es ablish a close examina ion o he UK economic impac be o e he o icial da a ge s eleased. 3.2. P edic ion Using AI This sec ion p esen s he me hodology used o economic p edic ions. Bo h con inuous and ca ego ical ime-se ies o ecas ing me hods we e ocused. A ime-se ies is de ined o be a collec ion o obse a ions made sequen ially h ough ime (Cha ield 2000). The concep o ime-se ies o ecas ing is o p edic u u e alues based on p e ious obse a ions. Based on he da a ype, o ecas ing can be di ided in o wo ca ego ies (con inuous and ca ego ical). Among hem, con inuous- alue o ecas ing is widely used in di e en domains such as me eo ology (Singh and Mohapa a 2019), epidemiology (Ben enu o e al. 2020), and economics. Following his endency, ini ially, we used a con inuous ime-se ies o ecas ing app oach o make u u e p edic ions, as de ailed in he sec ion ‘Con inuous Time-se ies Fo ecas ing’. Howe e , esul s ob ained by ini ial expe imen s demons a ed ha con inuous ime- se ies o ecas ing canno accu a ely cap u e he unexpec ed all o he economy due o COVID-19. I was an ex eme all compa ed o he p e ious alues, which a e used o ain he models. F om he pe spec i e o machine lea ning, models canno accu a ely p edic such unseen beha iou s. The e o e, we decided o conduc ca ego ical o ecas ing conside ing he g ow h and all o he economic measu es as desc ibed in he ‘Ca ego - ical Time-se ies Fo ecas ing’ sec ion. The g ow h o all o a pa icula ime poin was measu ed compa ed o a p e ious ime poin . The main pu pose o using ca ego ies is o con e unexpec ed a ia ions in o known o ma s. 3.2.1. Con inuous Time-Se ies Fo ecas ing Fo con inuous ime-se ies o ecas ing, we chose he Au o eg essi e In eg a ed Mo - ing A e age (ARIMA) model, conside ing i s popula i y among p e ious esea ch and simplici y 24 . Re e ing o ecen li e a u e, he ARIMA model was success ully used o p edic he epidemiological end o COVID-19 (Hewamalage e al. 2021), ood demand (Fa ah e al. 2018) , and g ow h in GDP. As a s a is ical model, ARIMA wo ks well wi h small da a se s, while a i icial neu al ne wo k-based models such as long sho - e m memo y (LSTM) need a huge amoun o da a o p ope lea ning. ARIMA model is a gene alised e sion o he Au o eg essi e Mo ing A e age (ARMA) model. As he name depic s, ARIMA combines he Au o eg ession (AR), Mo ing A e age (MA) model and a di e encing p ep ocessing s ep named in eg a ion (I) which makes he ime-se ies s a ione y. Based on hese h ee componen s, ARIMA equi es h ee hy- pe pa ame e s, he o de o AR (p) , he o de o di e encing (d) and he o de o MA (q) . Ma hema ically, ARIMA(p,d,q)model is o med using he equa ion: x =c+ p ∑ i=1 ∅ix −i+e + q ∑ j=0 θje −j whe e x a e ime-se ies alues, c is he cons an e m, ∅i is he coe icien o i h au o eg es- si e pa ame e , θj is he coe icien o j h mo ing a e age pa ame e and e a e esiduals o e o e ms (Fa ah e al. 2018). 3.2.2. Ca ego ical Time-Se ies Fo ecas ing We p opose a ca ego ical ime-se ies o ecas ing app oach o mi iga e he impac o unexpec ed changes on model p edic ions. Due o unexpec ed a ia ions, he e o a e be ween he ac ual and he p edic ed alues can be highly inc eased, and p edic ions can become less use ul. To p o ide use ul p edic ions in such si ua ions, we can use ca ego ical ime-se ies o ecas ing, since he limi a ion o possible alues by ca ego ies allows he con e sion o anomalous a ia ions in o known ep esen a ions. Economies 2021,9, 137 6 o 19 ARIMA-based Ca ego ical Fo ecas ing: As he ini ial app oach, we con e ed ac ual ime-se ies and p edic ions by he ARIMA model in o a se ies o ca ego ies conside ing consecu i e alue changes. A e he con e sion, we used a se o e alua ion me ics commonly used wi h classi ica ion asks o e alua e he model pe o mance om he aspec o ca ego ical p edic ions. The ca ego ies need o be de ined conside ing he a ge ed use g oup and in o ma ion equi emen s. Focusing on hese aspec s, we conside ed h ee ca ego ies ((1) g ow h, (2) cons an , and (3) all) in his esea ch. Each alue a ime is con e ed in o he ca ego ical o ma acco ding o he s a egies men ioned below. alue a > alue a −1→G ow h alue a = alue a −1→Cons an alue a < alue a −1→Fall SAX-VSM-based Ca ego ical Fo ecas ing: La e , we applied a classi ica ion p oce- du e named SAX-VSM which combines Symbolic Agg ega e app oXima ion (SAX) and Vec o Space Model (VSM) o make ca ego ical p edic ions. SAX was ound o be an e icien algo i hm o inding ime se ies disco ds (Keogh e al. 2005) and i was widely used wi h ime-se ies pa e n disco e y asks such as hea wa e e en de ec ion (He e a e al. 2016) and mic o-blog e en disco e y (S ilo and Vela di 2016). Since he ocused economic mea- su es ha e unexpec ed changes due o he pandemic si ua ion, his algo i hm’s abili ies can be e ec i ely used o make accu a e p edic ions. The VSM model is combined wi h SAX o acili a e ime-se ies classi ica ion. SAX con e s a ime se ies ( C ) o a w -dimensional space c=c1 , c2 , . . . cw using Piecewise Agg ega e App oxima ion (PAA) ollowing he equa ion: ci=w n n wi ∑ j=n w(i−1)+1 cj whe e cj is he j h elemen and n is he leng h o he o iginal se ies. Then each PAA ep e- sen a ion is con e ed in o a symbol o an alphabe o size a using a lookup able so ha he se ies is inally ep esen ed by a wo d which uses o u he analyses (Keogh e al. 2005) . SAX-VSM con e s a aining se in o a bag o wo ds pe class using a sliding window o leng h w0 o e he se ies and SAX algo i hm. In summa y, h ee hype -pa ame e s wo d size ( w ), alphabe size ( a ), and window size ( w0 ) a e equi ed by SAX-VSM. Using he gene a ed wo ds, a e m equency-in e se documen equency ( -id ) weigh ed ec o space is buil o he aining da a o acili a e he classi ica ion (Senin and Malinchik 2013). Gi en a se ies, we con e ed i o a se o subse ies wi h co esponding classes o use wi h his algo i hm. Simila o he abo e-men ioned app oach; ARIMA-based ca ego ical o ecas ing, we conside ed h ee ca ego ies ((1) g ow h, (2) cons an , and (3) all) as he classes in his app oach oo. Howe e , unlike he abo e scena io, compa isons be ween alues will no always happen be ween consecu i e alues, since he compa isons should use a leas one known alue du ing he da a p epa a ion. Fo example, i he ocus is o p edic he ca ego y o he alue a p succeeding ime poin s, ini ially he o iginal ime se ies need o be sepa a ed in o s -leng h subse ies as shown in Figu e 1 o p epa e aining da a. Then he alues in he las ime poin o each subse ies a e compa ed wi h he alues a p succeeding ime poin o gene a e classes. Fo subse ies 1 , class assignmen happens acco ding o he ollowing s a egies. Economies 2021,9, 137 7 o 19 Economies 2021, 9, x FOR PEER REVIEW 7 o 19 Figu e 1. SAX-VSM aining da a p epa a ion. Simila ly, classes a e assigned o o he subse ies and he a ge o he model buil using hese aining da a is o p edic he ca ego y o alue a 𝑝 succeeding ime poin compa ed o he las a ailable alue. I 𝑝=1, alues a consecu i e ime poin s and o h- e wise, alues a non-consecu i e ime poin s will be compa ed du ing ca ego y gene a- ion. Also, unlike he abo e-men ioned app oach, sepa a e SAX-VSM models need o be gene a ed pe each a ge ed 𝑝 alue. 4. Resul s This s udy included da a on GDP and h ee indus ies o e an ex ensi e pe iod o compa e he impac o COVID-19 wi h p e ious economic c ises obse ed in he ecen his o y o he UK economy in which eliable da a a e a ailable. When he pa e n o changes o GDP was analysed since 1947, which is he da a i s a ailable in ONS, we iden i ied signi ican d ops which we e ela ed o speci ic e en s. Howe e , he d op e- sul ed om COVID-19 is unp eceden ed (Figu e 2 and Table A1 in Appendix A). Appen- dix A ep esen s he mos signi ican economic ecessions in qua e ly GDP since eco ds began in 1955. Figu e 2. Pa e n o changes in GDP om 1955 un il he end o 2020 highligh ing he mos signi ican d ops and he pa e n o eco e y. 0.0 100,000.0 200,000.0 300,000.0 400,000.0 500,000.0 600,000.0 1955 Q1 1956 Q4 1958 Q3 1960 Q2 1962 Q1 1963 Q4 1965 Q3 1967 Q2 1969 Q1 1970 Q4 1972 Q3 1974 Q2 1976 Q1 1977 Q4 1979 Q3 1981 Q2 1983 Q1 1984 Q4 1986 Q3 1988 Q2 1990 Q1 1991 Q4 1993 Q3 1995 Q2 1997 Q1 1998 Q4 2000 Q3 2002 Q2 2004 Q1 2005 Q4 2007 Q3 2009 Q2 2011 Q1 2012 Q4 2014 Q3 2016 Q2 2018 Q1 2019 Q4 Chained olume measu es/£M Qua e G oss Domes ic P oduc (GDP) Figu e 1. SAX-VSM aining da a p epa a ion. xs+p>xs→G ow h xs+p=xs→Cons an xs+p<xs→Fall Simila ly, classes a e assigned o o he subse ies and he a ge o he model buil using hese aining da a is o p edic he ca ego y o alue a p succeeding ime poin compa ed o he las a ailable alue. I p= 1, alues a consecu i e ime poin s and o he wise, alues a non-consecu i e ime poin s will be compa ed du ing ca ego y gene a ion. Also, unlike he abo e-men ioned app oach, sepa a e SAX-VSM models need o be gene a ed pe each a ge ed p alue. 4. Resul s This s udy included da a on GDP and h ee indus ies o e an ex ensi e pe iod o compa e he impac o COVID-19 wi h p e ious economic c ises obse ed in he ecen his o y o he UK economy in which eliable da a a e a ailable. When he pa e n o changes o GDP was analysed since 1947, which is he da a i s a ailable in ONS, we iden i ied signi ican d ops which we e ela ed o speci ic e en s. Howe e , he d op esul ed om COVID-19 is unp eceden ed (Figu e 2and Table A1 in Appendix A). Appendix A ep esen s he mos signi ican economic ecessions in qua e ly GDP since eco ds began in 1955. Economies 2021, 9, x FOR PEER REVIEW 7 o 19 Figu e 1. SAX-VSM aining da a p epa a ion. Simila ly, classes a e assigned o o he subse ies and he a ge o he model buil using hese aining da a is o p edic he ca ego y o alue a 𝑝 succeeding ime poin compa ed o he las a ailable alue. I 𝑝=1, alues a consecu i e ime poin s and o h- e wise, alues a non-consecu i e ime poin s will be compa ed du ing ca ego y gene a- ion. Also, unlike he abo e-men ioned app oach, sepa a e SAX-VSM models need o be gene a ed pe each a ge ed 𝑝 alue. 4. Resul s This s udy included da a on GDP and h ee indus ies o e an ex ensi e pe iod o compa e he impac o COVID-19 wi h p e ious economic c ises obse ed in he ecen his o y o he UK economy in which eliable da a a e a ailable. When he pa e n o changes o GDP was analysed since 1947, which is he da a i s a ailable in ONS, we iden i ied signi ican d ops which we e ela ed o speci ic e en s. Howe e , he d op e- sul ed om COVID-19 is unp eceden ed (Figu e 2 and Table A1 in Appendix A). Appen- dix A ep esen s he mos signi ican economic ecessions in qua e ly GDP since eco ds began in 1955. Figu e 2. Pa e n o changes in GDP om 1955 un il he end o 2020 highligh ing he mos signi ican d ops and he pa e n o eco e y. 0.0 100,000.0 200,000.0 300,000.0 400,000.0 500,000.0 600,000.0 1955 Q1 1956 Q4 1958 Q3 1960 Q2 1962 Q1 1963 Q4 1965 Q3 1967 Q2 1969 Q1 1970 Q4 1972 Q3 1974 Q2 1976 Q1 1977 Q4 1979 Q3 1981 Q2 1983 Q1 1984 Q4 1986 Q3 1988 Q2 1990 Q1 1991 Q4 1993 Q3 1995 Q2 1997 Q1 1998 Q4 2000 Q3 2002 Q2 2004 Q1 2005 Q4 2007 Q3 2009 Q2 2011 Q1 2012 Q4 2014 Q3 2016 Q2 2018 Q1 2019 Q4 Chained olume measu es/£M Qua e G oss Domes ic P oduc (GDP) Figu e 2. Pa e n o changes in GDP om 1955 un il he end o 2020 highligh ing he mos signi ican d ops and he pa e n o eco e y. Economies 2021,9, 137 8 o 19 Table 1shows ha he decline in he qua e ly GDP ollowing he ini ial pe iod o he COVID-19 pandemic is se e e ( − 21%) compa ed o o he d ops in he qua e ly GDP o e he las 60 yea s in he UK (p< 0.0001). Table 1. Pe iods o majo ecessions in he UK wi h he pa e n o changes in he g oss domes ic p oduc (GDP). Pe iod Reason o he Down u n To al D op in GDP o e he Pe iod/% Lowes Figu e Qua e Following he S a o he C isis How Many Qua e s Did I Take o Re u n o Baseline 2019 Q4 o 2020 Q2 COVID-19 −21.2 Second qua e o 2020 Ongoing 2008 Q1 o 2009 Q2 The 2008 global inancial c isis −5.9 Fou h qua e o 2008 Six qua e s 1974 Q2 o 1975 Q3 1974 Mine s’ s ikes −5.4 Second qua e o 1975 Fi e qua e s 1979 Q2 o 1981 Q1 Du ing he 19800s economic down u n −5.3 Second qua e o 1980 Se en qua e s 1990 Q2 o 1991 Q3 The ea ly 1990 ecession −2.0 Thi d qua e o 1990 Fi e qua e s The second mos signi ican d op was obse ed ollowing he global inancial-economic c isis o 2008, and he in luenzas pandemic ha hi he UK a he same ime ( Ba o e al. 2020 ). Howe e , he GDP d op was >3.5 olds smalle han ha obse ed in COVID-19 ( − 5.9%) and i ook 4 ou qua e s o each he g ea es d op compa ed o wo qua e s in he COVID- 19 pandemic. The wo o he c ises in he ecen UK economic his o y we e obse ed in 1974 (Mine s’ s ikes) and 1980 (economic down u n) wi h GDP d ops o − 5.4% and − 5.3%, espec i ely. Du ing he i s ew mon hs o 1974, he e was a ou -week mine s’ s ike ac oss England. Du ing ha pe iod, he UK unemploymen inc eased, and in la ion hi 26% (Pa ing on 2020). No o he c ises we e associa ed wi h a d op g ea e han 5% as he ollowing decline iden i ied was − 2% ha was associa ed wi h he ea ly 1990 0 s ecession (Table 1). The ou pu le el o each indus y (se ice indus y, manu ac u ing indus y and cons uc ion indus y), can be seen in Figu e 3. Economies 2021, 9, x FOR PEER REVIEW 8 o 19 Table 1 shows ha he decline in he qua e ly GDP ollowing he ini ial pe iod o he COVID-19 pandemic is se e e (−21%) compa ed o o he d ops in he qua e ly GDP o e he las 60 yea s in he UK (p < 0.0001). Table 1. Pe iods o majo ecessions in he UK wi h he pa e n o changes in he g oss domes ic p oduc (GDP). Pe iod Reason o he Down u n To al D op in GDP o e he Pe iod/% Lowes Figu e Qua e Following he S a o he C isis How Many Qua e s Did I Take o Re u n o Baseline 2019 Q4 o 2020 Q2 COVID-19 −21.2 Second qua e o 2020 Ongoing 2008 Q1 o 2009 Q2 The 2008 global inancial c isis −5.9 Fou h qua e o 2008 Six qua e s 1974 Q2 o 1975 Q3 1974 Mine s’ s ikes −5.4 Second qua e o 1975 Fi e qua e s 1979 Q2 o 1981 Q1 Du ing he 1980′s economic down u n −5.3 Second qua e o 1980 Se en qua e s 1990 Q2 o 1991 Q3 The ea ly 1990 ecession −2.0 Thi d qua e o 1990 Fi e qua e s The second mos signi ican d op was obse ed ollowing he global inancial-eco- nomic c isis o 2008, and he in luenzas pandemic ha hi he UK a he same ime (Ba o e al. 2020). Howe e , he GDP d op was >3.5 olds smalle han ha obse ed in COVID- 19 (−5.9%) and i ook 4 ou qua e s o each he g ea es d op compa ed o wo qua e s in he COVID-19 pandemic. The wo o he c ises in he ecen UK economic his o y we e obse ed in 1974 (Mine s’ s ikes) and 1980 (economic down u n) wi h GDP d ops o −5.4% and −5.3%, espec i ely. Du ing he i s ew mon hs o 1974, he e was a ou -week mine s’ s ike ac oss England. Du ing ha pe iod, he UK unemploymen inc eased, and in la ion hi 26% (Pa ing on 2020). No o he c ises we e associa ed wi h a d op g ea e han 5% as he ollowing decline iden i ied was −2% ha was associa ed wi h he ea ly 1990′s ecession (Table 1). The ou pu le el o each indus y (se ice indus y, manu ac u - ing indus y and cons uc ion indus y), can be seen in Figu e 3. (a) 0.00 100,000.00 200,000.00 300,000.00 400,000.00 500,000.00 600,000.00 700,000.00 1997 Q1 1997 Q4 1998 Q3 1999 Q2 2000 Q1 2000 Q4 2001 Q3 2002 Q2 2003 Q1 2003 Q4 2004 Q3 2005 Q2 2006 Q1 2006 Q4 2007 Q3 2008 Q2 2009 Q1 2009 Q4 2010 Q3 2011 Q2 2012 Q1 2012 Q4 2013 Q3 2014 Q2 2015 Q1 2015 Q4 2016 Q3 2017 Q2 2018 Q1 2018 Q4 2019 Q3 2020 Q2 2021 Q1 Ou pu /£ Qua e Ou pu in he Se ice indus y Figu e 3. Con . Economies 2021,9, 137 15 o 19 Bank) 2021), while he obse ed decline was − 8.4%. The PwC g oup (Fo es e al. 2021) using he dynamic ac o model p edic ed GDP g ow h a e yea ly and mon hly. Howe e , hei p edic ion model showed a sligh di e ence om he obse ed igu es. Fo ins an , in Sep embe 2020 hey p edic ed a dec ease o − 14% o he hi d qua e o 2020 and ha he o e all dec ease in GDP would be be ween − 11% and − 12%. Howe e , he obse ed ONS da a indica ed ha he hi d-qua e GDP was +15.5% compa ed o he p e ious qua e and i was − 12.4% compa ed o he baseline qua e (Q4 2019) whe eas he yea ly GDP dec ease was − 8.4%. In 2020 Qua e 4, GDP showed an inc ease o 2.2%, which was signi ican ly highe han he p edic ed igu e o − 3.7% in he p e ious PwC economic epo (Fo es e al. 2021). The published a e age o ecas o 2021 GDP g ow h was om 3.7% o 4.3% (Ha a i e al. 2021). The a e age o ecas o 2021 Q1 GDP was a decline by 3.5% (HM T easu y 2021) o 4% (BoE (Bank o England) 2021) compa ed wi h he p e ious qua e based on he e ec o he lockdown. In his s udy, we used con inuous and ca ego ical ime-se ies o ecas ing-based ap- p oaches o make u u e p edic ions conside ing hei success ul applica ion in a wide ange o ields. Unlike mos s udies conduc ed in he a ea o economic analysis, we ocused on ca ego ical o ecas ing in ou app oach o mi iga e he impac o unexpec ed a ia ions on p edic ions. Ou p edic ions co e he u u e alues o GDP and o he indus y (se ice, manu ac u ing and cons uc ion) ou pu s in all he qua e s in 2021. In addi ion o he con inuous- alue p edic ions, we p edic ed he ca ego ical alues conside ing he ca e- go ies g ow h, cons an and all compa ed o he alues in he p e ious ime poin and he las qua e o 2020 o p o ide mo e insigh in o he u u e economy. Based on he p edic ions, GDP indica es a nea ly s eady-s a e wi h a g ow h o a ound 0.1% in all qua e s o 2021 wi h no signi ican inc ease o dec ease. This igu e seems di e en when i is compa ed o he p edic ed igu e o PwC using hei o ecas ing model anging om +3% o +7%, depending on hei quick o slow esponse scena ios. Fu he - mo e, when compa ing he p edic ions made o yea ly GDP g ow h compa ed o he inal qua e o 2019, hey ob ained a dec ease anging om − 6%– − 8%. This is compa able o he esul s we ha e ob ained om − 8.4% in GDP g ow h compa ed o he baseline using ou AI model. Ou p edic ions on GDP es ima e ha in he i s qua e o 2021 he e will be a dec ease o 0.3%. This can be expec ed as he hi d lockdown was en o ced du ing he i s qua e o he yea and businesses we e much mo e p epa ed han he p e ious lockdown o be able o s ill ope a e and h i e unde he new condi ions. Howe e , in he es o he qua e s o 2021, he e will be g ow h in GDP acco ding o he p edic ions. 6. Conclusions Mos s udies ha analysed he impac s o he COVID-19 on he UK economy ha e simila conclusions o ou s udy, emphasizing he ex ao dina y impac s on he di e en sec o s o he UK economy. The common line in he economic li e a u e is ep esen ed in he unp eceden ed impac s o he pandemic on he UK economy, which is associa ed wi h he mi iga ion s a egies ha he go e nmen imposes and is depending on how long hese s a egies will las . The cos s o hese s a egies a e due o los jobs because o lowe demand and business closu es. The esea ch pape s also e ealed a simila nega i e impac o he pandemic on he same sec o s we ocused on in ou s udy (cons uc ion, se ices, and p oduc ion)l The speci ic na u e o he UK makes i mo e likely o ace la ge economic impac s o he pandemic due o B exi which is likely o inc ease inpu cos s by inc easing ba ie s o ade be ween he UK and he EU. Fu he mo e, he UK is likely o ace a majo change in i s labou ma ke , which will equi e mo e suppo om he go e nmen o imp o e skills and cope wi h his change o e icien ly alloca e i s esou ces. The UK’s ins i u ional se ing may also ha e been a d i e o his poo pe o mance owa ds he pandemic, which leads o he p o ision o inadequa e in o ma ion ha a ec ed he e alua ion p ocess o he pandemic and he implemen a ion o he p ope policy. One o he impo an policies in dealing wi h he pandemic is he need o de elop e ec i e sys ems o pa icipa ion a all le els o human o ganiza ion. Economies 2021,9, 137 16 o 19 The pandemic ep esen s no only addi ional cos s on he UK’s economy bu also he challenges o he balance be ween he di e en economic goals o he go e nmen , inc easing jobs and p omo ing e iciency on he one hand and balancing be ween he bu den o he pandemic on he UK economy and he bu den o he pos -pandemic cos s on he o he . Imposing mi iga ion s a egies would lowe dea h a es and igh he pandemic, bu i imposes much highe cos s, in he long un, cos s ha would ep esen e en much mo e challenges o he policymake in he u u e. In sho , he case o he UK economy wi h i s na u e acing he pandemic c ea es a dilemma ha ep esen s a majo challenge o any decision make . S udying he pa e n o eco e y ollowing p e ious economic c ises may help o accu a ely o ecas he pa e n o change in he u u e. In addi ion, s udying he impac s o he c isis u ilising a su icien ime pe iod ollowing he pandemic is likely o p o ide mo e accu a e igu es on he long- e m changes in he economy as he e y ea ly igu es may be misleading and could exagge a e he impac s. AI p o ides an impo an ool o la ge scale analysis o economic da a and can p o ide he obus ool o mul ipa ame e and mul idimensional da a analysis o e alua ing he impac s and p edic ing u u e pa e ns. The o ecas s p o ided in he cu en s udy can help make be e decisions, de e mine mone a y and iscal policies, and could be used by he go e nmen and business o help de e mine hei u u e s a egy, budge s and mul i-yea plans. Howe e , u he alida ion o AI models is wa an ed o ensu e mo e eliabili y and u ili y o his app oach. The mo e a iables, e en s, and ime poin s a e used in he model, he mo e accu a e he p edic ion. Au ho Con ibu ions: Concep ualiza ion, A.R., M.M.G., E.R. and M.M.A.; me hodology, H.H., M.M.G. and M.M.A.; so wa e, H.H.; alida ion, A.R. and H.H.; o mal analysis, A.R. and H.H.; in es iga ion, A.R. and H.H.; esou ces, A.R.; w i ing—o iginal d a p epa a ion, A.R. and H.H.; w i ing— e iew and edi ing, A.R., H.H., D.R., M.M.G., E.R. and M.M.A.; isualiza ion, A.R. and H.H.; supe ision, M.M.A.; p ojec adminis a ion, M.M.A.; All au ho s ha e ead and ag eed o he published e sion o he manusc ip . Funding: This esea ch ecei ed no ex e nal unding. Ins i u ional Re iew Boa d S a emen : No applicable. In o med Consen S a emen : No applicable. Da a A ailabili y S a emen : The da a used in his s udy we e collec ed om he O ice o Na ional S a is ics (ONS 2021) and he implemen a ion code is a ailable online a h ps://gi hub.com/HHansi/ Economic-P edic ion. Con lic s o In e es : The au ho s decla e no con lic o in e es . Appendix A Table A1. The main eigh ecessions in qua e ly GDP obse ed since eco ds began in 1955. Fi s Qua e Las Qua e Numbe o Qua e s o Nega i e G ow h To al Decline/% 1956 Q2 1956 Q3 Two qua e s −0.3 1961 Q3 1961 Q4 Two qua e s −0.7 1973 Q3 1974 Q1 Th ee qua e s −4.1 1975 Q2 1975 Q3 Two qua e s −2.0 1980 Q1 1981 Q1 Fi e qua e s −4.2 1990 Q3 1991 Q3 Fi e qua e s −2.0 2008 Q2 2009 Q2 Fi e qua e s −6.0 2020 Q1 2020 Q2 (Keogh-B own e al. 2020) Two qua e s −22.1 Economies 2021,9, 137 17 o 19 Appendix B The cons uc ion indus y was he mos a ec ed indus y showing a dec ease o 35%. This is he la ges qua e ly all in he cons uc ion indus y’s GDP since eco ds began, mo e han ou imes bigge han he second-la ges all in he cons uc ion indus y. The in luenzas pandemic o ins ance esul ed in a educ ion in he GDP by − 2.25% o − 5.2% (Keogh-B own e al. 2020;McKibbin and Fe nando 2020). This s udy also showed ha all new in es men s in he cons uc ion indus y ha e dec eased by 11% since Feb ua y 2020, an index o 473,444 was ob ained o he igu es in July. The mos signi ican d op and he lowes alue ob ained we e in Ap il, jus ollowing he in oduc ion o he i s lockdown; he index d opped by 35.6% om 503,234 o 298,961. Following on om Ap il, he ou pu le els o he indus y as a whole s a ed o ise, as businesses began o adap o he si ua ion. The sec o ha was he mos signi ican ly a ec ed was public housing. This sec o ’s ou pu le els d opped by 67% in Ap il 2020, as he majo i y o public cons uc ion wo k on new housing was hal ed (Boissay and Rungcha oenki kul 2020). Keogh-B own e al. (2020) iden i ied a educ ion in he consume indus y ou pu le el du ing he in luenzas ou b eak, which was 9% highe han he index ob ained in his s udy. The only sec o o he cons uc ion indus y ha achie ed a highe index was he public in as uc u e sec o , achie ing an index ha was 6.1% highe han be o e he lockdown. This sec o was leas a ec ed by he lockdown o e all. The p oduc ion indus y was a ec ed bu no as badly a ec ed as he cons uc ion indus y. The second qua e o 2020 GDP o he p oduc ion indus y was − 20%. The second-la ges qua e ly all o he p oduc ion indus y was obse ed in he i s qua e o 2009. Howe e , he all in GDP was 5%, Fou imes smalle han he dec ease in GDP in he second qua e o 2020. This is u he e idence ha each indus y was a ec ed di e en ly by he pandemic, which ep esen s ha he p oduc ion indus y was he mos a ec ed indus y by he lockdown pe iod, bu no o he ex en ha he cons uc ion indus y’s GDP had d opped. As highligh ed abo e, he absolu e educ ion in he h ee ca ego ies indica ed ha cons uc ion was a ec ed he mos wi h a 35% educ ion compa ed o he se ice indus y and p oduc ion indus y (19% and 20%). Howe e , when analysing he ela i e educ ion as compa ed o he p e ious alls in hese indus y pe o mances du ing he ecen economic his o y in he UK, we ound ha he ela i e educ ion in he se ice indus y was he highes wi h mo e han 24 olds compa ed o in luenzas ou b eak which a ec ed he UK in 2009. A he same ime, he d ops in cons uc ion and se ice we e i e imes and ou imes when compa ed o he 2009 economic hi , espec i ely. In e es ingly in he 2009 i us ou b eak, he bigges impac was obse ed in he cons uc ion and p oduc ion indus y, and he leas a ec ed indus y was he se ice wi h a ela i e educ ion in se ice was nine imes and six imes less compa ed o he cons uc ion and p oduc ion indus y. Howe e , du ing he COVID-19 c isis, he ela i e educ ion in he se ice indus y was h ee imes as big as he cons uc ion indus y and was wo imes as big as he p oduc ion indus y. These changes in he pa e ns o d ops o a ious indus ies a e likely o be a e lec ion in he scale o he p e en i e measu es aken in bo h ou b eaks and he global na u e o he COVID-19 compa ed o he limi ed e ec o he in luence ou b eak in 2009. The se ice indus y GDP saw he smalles decline o all h ee indus ies. In he second qua e o 2020, he GDP ell by 19% because o he lockdown imposed in Ma ch. Tha is nea ly en imes he alue o he second-la ges dec ease in he se ice indus y GDP o − 2.3%, which was obse ed in he i s qua e o 2020. The 2009 inancial c isis and he in luenza ou b eak caused a dec ease in he i s qua e by 0.8%. This dec ease is 96% smalle han he decline obse ed due o he lockdown imposed. This is u he e idence ha each indus y was a ec ed di e en ly by he pandemic and ep esen ed ha he se ice indus y was he leas a ec ed indus y by he lockdown pe iod. This indica es ha he se ice indus y was a ec ed signi ican ly, bu lowe han he cons uc ion indus y, as a esul o he na ional lockdown. Economies 2021,9, 137 18 o 19 Re e ences Ahuja, Abhimanyu S. 2019. The impac o a i icial in elligence in medicine on he u u e ole o he physician. Pee J 7: e7702. [C ossRe ] [PubMed] A slan, Janan, and Ku K. Benke. 2021. A i icial In elligence and Teleheal h may p o ide ea ly wa ning o epidemics. F on ie s in A i icial In elligence 4: 556848. [C ossRe ] Balm o d, Ben, James D. Annan, Julia C. Ha g ea es, Ma ina Al oè, and Ian J. Ba eman. 2020. C oss-coun y compa isons o COVID-19: Policy, poli ics and he p ice o li e. En i onmen al and Resou ce Economics 76: 525–51. [C ossRe ] [PubMed] Ba o, Robe J., JoséF. U súa, and Joanna Weng. 2020. The Co ona i us and he G ea In luenza Pandemic: Lessons om he “Spanish Flu” o he Co ona i us’s Po en ial E ec s on Mo ali y and Economic Ac i i y (No. w26866). Camb idge: Na ional Bu eau o Economic Resea ch. Becke , Andy. 2020. Nobody Denies Johnson’s Go e nmen is Incompe en . Bu Do Enough Vo e s Ca e? The Gua dian. Augus 22. A ailable online: h ps://www. hegua dian.com/commen is ee/2020/aug/22/bo is-johnson-go e nmen -incompe en - p ime-minis e -compe en (accessed on 28 Sep embe 2020). Ben enu o, Domenico, Ma a Gio ane i, Lazza o Vassallo, Sil ia Angele i, and Massimo Ciccozzi. 2020. Applica ion o he ARIMA model on he COVID-2019 epidemic da ase . Da a in B ie 29: 105340. [C ossRe ] [PubMed] BFPG. 2020. The B i ish Fo eign Policy G oup. A ailable online: h ps://b pg.co.uk/2020/04/co id-19- imeline/ (accessed on 1 Oc obe 2020). BoE (Bank o England). 2021. Mone a y Policy Repo . London: Bank o England, p. 2. Boissay, F ede ic, and Phu ichai Rungcha oenki kul. 2020. Mac oeconomic E ec s o COVID-19: An Ea ly Re iew (No. 7). Basel: Bank o In e na ional Se lemen s. Cha ield, Ch is. 2000. Time-Se ies Fo ecas ing. Boca Ra on: CRC P ess. Child, John. 2021. O ganiza ional pa icipa ion in pos -co id socie y—I s con ibu ions and enabling condi ions. In e na ional Re iew o Applied Economics 35: 117–46. [C ossRe ] Coakley, Je y, Pe iklis Kougoulis, and John C. Nanke is. 2014. Como emen and FTSE 100 index changes. In e na ional Jou nal o Beha iou al Accoun ing and Finance 4: 93–112. [C ossRe ] Cos a Dias, Monica, Robe Joyce, Fabien Pos el-Vinay, and Xiaowei Xu. 2020. The challenges o labou ma ke policy du ing he COVID-19 pandemic. Fiscal S udies 41: 371–82. [C ossRe ] De Lyon, Josh, and Swa i Dhing a. 2021. The Impac s o COVID-19 and B exi on he UK Economy: Ea ly E idence in 2021. London: Cen e o Economic Pe o mance, London School o Economics and Poli ical Science. Fa ah, Jamal, La i a Ezzine, Zanab Aman, Haj El Moussami, and Abdeslam Lachhab. 2018. Fo ecas ing o demand using ARIMA model. In e na ional Jou nal o Enginee ing Business Managemen 10: 1847979018808673. [C ossRe ] Fo es , Nick, Hannah Audino, Jona han Gillham, and Ba e Kupelian. 2021. UK Economic Upda e. © P iceWa e house Coope s LLP (PwC Ne wo k). A ailable online: h ps://www.pwc.co.uk/ (accessed on 1 Ma ch 2021). Fu, Xin, Hao Yang, Chenxi Liu, Jianwei Wang, and Yinhai Wang. 2019. A hyb id neu al ne wo k o la ge-scale exp essway ne wo k OD p edic ion based on oll da a. PLoS ONE 14: e0217241. [C ossRe ] G an z, Ky a H., Madhu a S Rane, Hen ik Salje, G ego y E. Glass, S ephen E. Schach e le, and De ek A.T. Cummings. 2016. Dispa i ies in in luenza mo ali y and ansmission ela ed o sociodemog aphic ac o s wi hin Chicago in he pandemic o 1918. P oceedings o he Na ional Academy o Sciences o he Uni ed S a es o Ame ica 113: 13839–44. [C ossRe ] G i i h, Rachel, Pe e Le ell, and Rebekah S oud. 2020. The Impac o COVID-19 on Sha e P ices in he UK. Fiscal S udies 41: 363–69. [C ossRe ] [PubMed] Ha a i, Daniel, Ma hew Keep, and Philip B ien. 2021. Co ona i us: Economic Impac . B ie ing Pape . Numbe 8866, 26 Feb ua y 2021. London: The House o Commons Lib a y. A ailable online: h ps://commonslib a y.pa liamen .uk/ esea ch-b ie ings/cbp-8866 /(accessed on 24 Sep embe 2021). He e a, Manuel, Aida A Fe ei a, Da id Coley, and Ronaldo R.B. de Aquino. 2016. SAX-quan ile based mul i esolu ion app oach o inding hea wa e e en s in summe empe a u e ime se ies. AI Communica ions 29: 725–32. [C ossRe ] Hewamalage, Hansika, Ch is oph Be gmei , and Kasun Banda a. 2021. Recu en neu al ne wo ks o ime se ies o ecas ing: Cu en s a us and u u e di ec ions. In e na ional Jou nal o Fo ecas ing 37: 388–427. [C ossRe ] HM T easu y. 2021. Su ey o Independen Fo ecas s o he UK Economy: Feb ua y 2021. London: HM T easu y. Huang, Chaolin, Yeming Wang, Xingwang Li, Lili Ren, Jianping Zhao, Zhang Hu, Fan Li, Guohui Yi, Jiuyang Xu, and Xiaoying Gu. 2020. Clinical ea u es o pa ien s in ec ed wi h 2019 no el co ona i us in Wuhan, China. Lance 395: 497–506. [C ossRe ] Ib ahim, Asmaa, Paul Gamble, Ronnachai Ja oens i, Mohammed M Abdelsamea, C aig H. Me mel, Po-Hsuan Came on Chen, and Emad A. Rakha. 2020. A i icial in elligence in digi al b eas pa hology: Techniques and applica ions. B eas 49: 267–73. [C ossRe ] [PubMed] Kelso, Paul. 2020. Co ona i us: Economic Impac Bigge han Fi s Fea ed, Wa ns IMF. Sky News. A ailable online: h ps://news.sky. com/ (accessed on 10 Oc obe 2020). Keogh, Eamonn, Jessica Lin, and Ada Fu. 2005. Ho sax: E icien ly inding he mos unusual ime se ies subsequence. In P oceedings o he Fi h IEEE In e na ional Con e ence on Da a Mining (ICDM’05), Hous on, TX, USA, No embe 27–30; p. 8. Economies 2021,9, 137 19 o 19 Keogh-B own, Ma cus R., He ning Ta p Jensen, John W. Edmunds, and Richa d D. Smi h. 2020. The impac o COVID-19, associa ed beha iou s and policies on he UK economy: A compu able gene al equilib ium model. SSM-Popula ion Heal h 12: 100651. [C ossRe ] Kim, Myung Hwa, Ju Hyung Kim, Kyoungjin Lee, and Gwang-Yong Gim. 2021. The P edic ion o COVID-19 Using LSTM Algo i hms. In e na ional Jou nal o Ne wo ked and Dis ibu ed Compu ing 9: 19–24. [C ossRe ] McKibbin, Wa wick, and Roshen Fe nando. 2020. The Global Mac oeconomic Impac s o COVID-19: Se en Scena ios. Cen e o Applied Mac oeconomic Analysis (CAMA) Wo king Pape 19: 1–43. [C ossRe ] O’Donnell, Gus, and Ha y Begg. 2020. Fa om Well: The UK since COVID-19, and Lea ning o Follow he Science (s). Fiscal S udies 41: 761–804. [C ossRe ] ONS. 2021. O ice o Na ional S a is ics: Co ona i us and he Impac on Ou pu in he UK Economy. A ailable online: h ps://www. ons.go .uk/economy/g ossdomes icp oduc gdp/a icles/co ona i usand heimpac onou pu in heukeconomy/decembe 2020 (accessed on 20 Ma ch 2021). Paisley, Anne. 2020. COVID-19: The S a e o Global Pandemic P epa edness. A ailable online: h ps://www.one.o g/in e na ional/ blog/s a e-o -global-pandemic-p epa edness/ (accessed on 1 Oc obe 2020). Pa ing on, Richa d. 2020. How Does UK’s COVID Recession Compa e wi h P e ious Ones? The Gua dian. Augus 12. A ailable online: h ps://www. hegua dian.com/business/2020/aug/12/how-does-uks-co id- ecession-compa e-wi h-p e ious-ones (accessed on 1 Oc obe 2020). Sake , Lance, Kelley Lee, Ba ba a Canni o, Anna Gilmo e, and Dia mid Campbell-Lend um. 2004. Globaliza ion and In ec ious Diseases: A Re iew o he Linkages. Gene a: UNDP/Wo ld Bank/WHO Special P og amme on T opical Diseases Resea ch. Senin, Pa el, and Se gey Malinchik. 2013. Sax- sm: In e p e able ime se ies classi ica ion using sax and ec o space model. In P oceedings o he 2013 IEEE 13 h In e na ional Con e ence on Da a Mining, Dallas, TX, USA, Decembe 7–10; pp. 1175–1180. Singh, S. N., and Abheejee Mohapa a. 2019. Repea ed wa ele ans o m based ARIMA model o e y sho - e m wind speed o ecas ing. Renewable Ene gy 136: 758–68. Song, Ligang, and Yixiao Zhou. 2020. The COVID-19 pandemic and i s impac on he global economy: Wha does i ake o u n c isis in o oppo uni y? China and Wo ld Economy 28: 1–25. [C ossRe ] S ephens, Ma k, Sian C oss, and Ga e h Luckwell. 2020. Co ona i us and he Impac on Ou pu in he UK Economy: June 2020; London: O ice o Na ional S a is ics, p. 12. S ilo, Gio anni, and Paola Vela di. 2016. E icien empo al mining o mic o-blog ex s and i s applica ion o e en disco e y. Da a Mining and Knowledge Disco e y 30: 372–402. [C ossRe ] TWB (The Wo ld Bank). 2021. Global Economic P ospec s 2021. A ailable online: h ps://www.wo ldbank.o g/en/publica ion/ global-economic-p ospec s (accessed on 24 Sep embe 2021). Wahl, B ian, Aline Cossy-Gan ne , S e an Ge mann, and Nina R. Schwalbe. 2018. A i icial in elligence (AI) and global heal h: How can AI con ibu e o heal h in esou ce-poo se ings? BMJ Global Heal h 3: e000798. [C ossRe ] [PubMed]