Usporedba performansi algoritama strojnog učenja u cilju vrednovanja financijske imovine na američkom tržištu kapitala

master's thesis

• Title (english): Performance comparison of machine learning algorithms for pricing financial assets in the US capital market
• Author: Mate Pavelić
• Mentor: Petra Posedel Šimović (mentor)
• Committee member: Zvonko Kostanjčar (član povjerenstva)
• Committee member: Stjepan Šebek (član povjerenstva)
• Committee member: Petra Posedel Šimović (predsjednik povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2021-07-13, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Electrical Engineering

Abstract

Algoritmi strojnog učenja moćan su alat u financijama za izradu predikcija kretanja cijena dionica i konstrukciju portfelja. U ovome radu napravljena je analiza tri algoritma primijenjenih na povijesne podatke s američkog dioničkog tržišta u svrhu određivanja budućih cijena dionica. Korišteni su OLS, LASSO i Random Forest algoritmi. U predikcijskim performansama kao i u konstrukciji portfelja dominira algoritam Random Forest. Konstruirane su dvije vrste portfelja: long-short i long-only.

Abstract (english)

Machine learning algorithms are a powerfull tool to use in finance to make predictions for future stock price movements and portfolio construction. In this paper we conduct an analysis of three algorithms applied on historic dana from american stock market in order to predict future stock prices. Used algorithms were OLS, LASSO and Random Forest. Regarding predictive performance as well as portfolio construction, Random Forest algorithm was shown to be the best. Two types of portfolios were constructed: long-short and long-only.

• Keywords: strojno učenje; vremenski niz; OLS; LASSO; Random Forest; portfelj; Sharpe omjer
• Keywords (english): machine learning; time series; OLS; LASSO; Random Forest; portfolio; Sharpe ratio
• Language: croatian
• URN:NBN: urn:nbn:hr:168:481615
• Study programme: Title: Information and Communication Technology; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka informacijske i komunikacijske tehnologije (magistar/magistra inženjer/inženjerka informacijske i komunikacijske tehnologije)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-01-13 09:18:27