Predviđanje volatilnosti financijskih vremenskih nizova

undergraduate thesis

• Title (english): Financial Time Series Volatility Forecasting
• Author: Jura Milković
• Mentor: Zvonko Kostanjčar (mentor)
• Committee member: Zvonko Kostanjčar (predsjednik povjerenstva)
• Committee member: Tomislav Kovačević (član povjerenstva)
• Committee member: Andro Merćep (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2022-07-15, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Zbog uske povezanosti volatilnosti s rizikom ulaganja, javlja se potreba za njenim predviđanjem. Prethodno se pokazalo da su modeli strojnog učenja uspješniji od standardnih statističkih pristupa u brojnim znanstvenim disciplinama, što je motiviralo istraživanje modela strojno učenja za predviđanje volatilnosti. U ovom radu, isprobani su linearni modeli i modeli slučajnih šuma za predviđanje volatilnosti definirane na tri različita načina. Modeli su trenirani i testirani na povijesnim podatcima tri različite vrste financijske imovine: dionica, kriptovaluta i valutni par. Eksperimenti su pokazali da je najmanja srednja kvadratna pogreška postignuta za predviđanje varijance eksponencijalno otežanih povrata koristeći modele slučajnih šuma.

Abstract (english)

Since volatility is closely related to investment risk, there is a need for its prediction. Machine learning models have already proven to be more successful than standard statistical approaches in several scientific disciplines, which has motivated the exploration of machine learning models for volatility prediction. This paper tests linear models and random forest models for volatility prediction defined in three different ways. The models were trained and tested on historical data from three different types of financial assets: Stocks, cryptocurrencies, and currency pairs. The experiments showed that the smallest mean square error in predicting the variance of exponentially weighted returns is achieved with random forest models.

• Keywords: volatilnost; strojno učenje; regresija
• Keywords (english): volatility; machine learning; regression
• Language: croatian
• URN:NBN: urn:nbn:hr:168:576099
• Study programme: Title: Computing; Study programme type: university; Study level: undergraduate; Academic / professional title: sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva (sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-01-15 10:52:32