Stock price prediction using neural networks within Weka software : a comparative analysis of two time periods

diplomski rad

• Autor: Filip Javor
• Mentor: Mirjana Pejić Bach (mentor)
• Član povjerenstva: Mario Spremić (član povjerenstva)
• Član povjerenstva: Božidar Jaković (član povjerenstva)
• Ustanova koja je dodijelila akademski / stručni stupanj: Sveučilište u Zagrebu; Ekonomski fakultet; Zagreb
• Datum i država obrane: 2023-10-02, Hrvatska
• Znanstveno / umjetničko područje, polje i grana: DRUŠTVENE ZNANOSTI; Ekonomija; Poslovna informatika

Sažetak

The main goal of this research thesis was to construct and test the effectiveness of a stock price prediction model based on neural networks, on two separate time periods, and determine whether the COVID-19 crisis had an impact on the stock market's predictability. Pre-covid and post-covid periods are the first and second, respectively. The official website of the New York Stock Exchange was used to gather the data for the research study. The pre-processed data was then loaded into the Weka program, where it was used to carry out the study. Furthermore, the software was used to create and assess the model. Once the model was initially designed, it was tested and improved upon, until satisfactory results were obtained. Prior to the research description and findings, an introduction to stock market prediction models, neural networks, and some limitations are addressed. Finally, the results and key findings are analyzed and discussed. The paper ends with a comprehensive conclusion.

• Ključne riječi: stock market prediction; neural networks; artificial intelligence
• Jezik: engleski
• URN:NBN: urn:nbn:hr:148:833151
• Studijski program: Naziv: Master Degree in Business; smjerovi: Marketing na engleskom jeziku, Managerial Informatics, Management, Marketing, Trade and International Business, Leadership; Smjer: Managerial Informatics; Vrsta studija: sveučilišni; Stupanj studija: diplomski; Akademski / stručni naziv: sveučilišni magistar ekonomije/sveučilišna magistra ekonomije (univ.mag.oec.)
• Vrsta resursa: Tekst
• Način izrade datoteke: Izvorno digitalna
• Prava pristupa: Pristup svim korisnicima iz sustava znanosti i visokog obrazovanja RH
• Datum i vrijeme pohrane: 2023-10-02 11:57:00