Otkrivanje anomalija u vremenskim nizovima korištenjem varijacijskog autoenkodera

master's thesis

• Title (english): Time Series Anomaly Detection Using Variational Autoencoder
• Author: Jozo Papić
• Mentor: Zvonko Kostanjčar (mentor)
• Committee member: Zvonko Kostanjčar (predsjednik povjerenstva)
• Committee member: Marko Subašić (član povjerenstva)
• Committee member: Stjepan Begušić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2021-07-14, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Rad se bavi primjenom metoda dubokog učenja za otkrivanje anomalije u vremenskim nizovima. Razrađen je teorijski opis modela varijacijskog autoenkodera i odrađeno je njegovo uvježbavanje. Uvježbani model varijacijskog autoenkodera se koristio nad vremenskim nizovima na financijskom tržištu te su u tu svrhu razrađene tri strategije trgovanja nad dionicama koje koriste informaciju sa izlaza varijacijskog autoenkodera.

Abstract (english)

This thesis shows the use of deep learning methods for anomaly detection on time series. For that purpose, a theoretical model of variational autoencoder has been made and trained. The trained variational autoencoder model has been used on time series on financial markets and for this case three trading strategies that use information from variational autoencoder output have been elaborated.

• Keywords: Algoritmi trgovanja; dionice; autoenkoderi; duboko učenje
• Keywords (english): Trading algorithms; stocks; autoencoders; deep learning
• Language: croatian
• URN:NBN: urn:nbn:hr:168:654754
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-01-13 15:47:07