kth.sePublications KTH
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Comparing machine learning models for predicting stock market volatility using social media sentiment: A comparison of the predictive power of the Artificial Neural Network, Support Vector Machine and Decision Trees models on price volatility using social media sentiment
KTH, School of Electrical Engineering and Computer Science (EECS).
KTH, School of Electrical Engineering and Computer Science (EECS).
2021 (English)Independent thesis Basic level (degree of Bachelor), 10 credits / 15 HE creditsStudent thesisAlternative title
Komparativ analys av sentimentbaserade maskininlärningsmodeller för förutsägelse av volatilitet på aktiemarknaden (Swedish)
Abstract [en]

We aimed to explore how the machine learning models Artificial Neural Network (ANN), Support Vector Machine (SVM) and Decision tree (DT) compared in analyzing the effects of investor sentiment (from the forum www.reddit.com/r/wallstreetbets) in conjunction with other key parameters, to predict asset price volatility of major US corporations. The paper explores the effect on asset price volatility that the addition of sentiment based indicators had on companies listed in the S&P500 index since 2012. None of the models we used could accurately predict the volatility of stocks using our collected sentiment and financial data. While social media sentiment has been shown by previous research to impact parts of financial markets, the market as a whole does not seem to be as susceptible to this interference as some analysts have suggested. Therefore, the training data for the algorithms had too much noise to find strong relationship. Furthermore we believe that more research is required in order to better understand which financial (or other) indicators play a role in shaping online sentiment. 

Abstract [sv]

I den här rapporten ämnade vi att undersöka hur väl maskininlärningsmodellerna Artificial Neural Network (ANN), Support Vector Machine (SVM) och Decision trees (DT) kan förutspå prisvolatilitet av enskilda aktier med hjälp av historiskt användarsentiment och finansiella nyckelindikatorer. Vi har använt det historiska sentimentet på forumet www.reddit.com/r/wallstreetbets och historisk prisdata för det Amerikanska börsindexet S&P500. Ingen modell kunde exakt förutspå prisvolatilitet av enskilda med hjälp av sentiment och den finansiella datan. Det indikerar att även om enskilda företag kan påverkas starkt så har marknaden som helhet inte påverkats i signifikant grad av den sentimentkälla vi undersökt. Därför hade träningsdatan för algoritmerna för mycket brus för att hitta starka relationer i datan. Vidare anser vi att mer forskning krävs för att etablera en länk mellan vilka finansella (eller andra) nyckeltal som påverkar sentimentet på social media.

Place, publisher, year, edition, pages
2021. , p. 40
Series
TRITA-EECS-EX ; 2021:661
National Category
Computer Sciences
Identifiers
URN: urn:nbn:se:kth:diva-304660OAI: oai:DiVA.org:kth-304660DiVA, id: diva2:1609945
Subject / course
Computer Science
Educational program
Master of Science in Engineering - Computer Science and Technology
Supervisors
Examiners
Available from: 2021-11-10 Created: 2021-11-09 Last updated: 2022-06-25Bibliographically approved

Open Access in DiVA

fulltext(1374 kB)1742 downloads
File information
File name FULLTEXT01.pdfFile size 1374 kBChecksum SHA-512
d04f302aaf43cacc2b7b85c47fc8711029c73ed33b04e277b77e2948fc2cccd628b9ea186977aac77f9a03c052ddbf734d58613cabf723e0dbf86267d38b3369
Type fulltextMimetype application/pdf

By organisation
School of Electrical Engineering and Computer Science (EECS)
Computer Sciences

Search outside of DiVA

GoogleGoogle Scholar
Total: 1743 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

urn-nbn

Altmetric score

urn-nbn
Total: 565 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf