Twitter predicts Naspers share price

Simple item page
dc.contributor.advisorConradie, Willieen_ZA
dc.contributor.authorSebitlo, Kgomotso Julianen_ZA
dc.contributor.otherStellenbosch University. Faculty of Economic and Management Sciences. Dept. of Statistics and Actuarial Science.en_ZA
dc.date.accessioned2022-03-07T07:19:19Z
dc.date.accessioned2022-04-29T09:24:12Z
dc.date.available2022-03-07T07:19:19Z
dc.date.available2022-04-29T09:24:12Z
dc.date.issued2022-04
dc.descriptionThesis (MCom)--Stellenbosch University, 2022.en_ZA
dc.description.abstractENGLISH SUMMARY: Naspers is a publicly listed South African company, and @Fin24 is the financial news Twitter account that belongs to Naspers. Using Natural Language Processing techniques, Twitter sentiment and emotional analysis is applied to extract emotions from @Fin24 and these emotions are used to predict the Naspers share price. Two Twitter emotion extraction methods are compared. The first method is a character-based recurrent neural network which was developed to specifically recognise emotions on Twitter, while the second method is lexicon-based and was developed to recognise emotions from any text. The first method emotion scores, particularly the two-day lag Anticipation score, is found to contain useful information that may assist in predicting the daily Naspers share price.en_ZA
dc.description.abstractAFRIKAANSE OPSOMMING: Naspers is ’n publiek-genoteerde Suid-Afrikaanse maatskappy, en @Fin24 is die finansi¨ele nuus Twitter-rekening wat aan Naspers behoort. Met behulp van Natuurlike Taal Prossesering tegnieke, Twitter-sentiment en emosionele analise toegepas om emosies uit @Fin24 te onttrek en hierdie emosies word gebruik om die Naspers-aandeelprys te voorspel. Twee Twitter-emosieonttrekkingsalgoritmes word vergelyk. Die eerste algoritme is ’n karakter-gebaseerde herhalende neurale netwerk wat ontwikkel is om spesifiek emosies op Twitter te herken, terwyl die tweede algoritme leksikon-gebaseer is en ontwikkel is om emosies uit enige teks te herken. Daar word gevind dat die eerste algoritme-emosietellings, veral die twee-dae vertraging Antisipasie-telling, nuttige inligting bevat wat kan help om die daaglikse Naspers-aandeelprys te voorspel.af_ZA
dc.description.versionMasters
dc.format.extentxvii, 116 pages : illustrations, includes annexures
dc.identifier.urihttp://hdl.handle.net/10019.1/124646
dc.language.isoen_ZAen_ZA
dc.publisherStellenbosch : Stellenbosch University
dc.rights.holderStellenbosch University
dc.subjectArtificial intelligenceen_ZA
dc.subjectSocial media -- Data processingen_ZA
dc.subjectMachine learningen_ZA
dc.subjectComputational emotion detectionen_ZA
dc.subjectUCTD
dc.titleTwitter predicts Naspers share priceen_ZA
dc.typeThesisen_ZA
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
sebitlo_twitter_2022.pdf
Size:
4.41 MB
Format:
Adobe Portable Document Format
Description:
Download
Collections
Masters Degrees (Statistics and Actuarial Science)