Predviđanje cijena kriptovaluta korištenjem analize vremenskih serija

master's thesis

• Title (english): Cryptocurrency Prediction Using Time Series Analysis
• Author: Toni Matić
• Mentor: Marin Šilić (mentor)
• Committee member: Marin Šilić (predsjednik povjerenstva)
• Committee member: Goran Delač (član povjerenstva)
• Committee member: Klemo Vladimir (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2023-07-04, Croatia

Abstract

Kriptovalute su digitalni oblik valuta kojima upravlja decentralizirano tržište. Ulagačima je cilj predvidjeti cijene kriptovaluta te tako profitirati. Cijene je kriptovaluta moguće gledati kao vremensku seriju, pa ih tako i analizirati ih pomoću metoda i alata za analizu vremenskih serija. Analiza vremenskih serija područje je dubinske analize podataka koje analizira serije podataka, a cilj joj je pronaći zanimljive obrasce te ih onda iskoristiti za druge korisne zadatke. Cilj je ovoga rada analizom vremenskih serija predvidjeti buduće cijene kriptovaluta metodama strojnog učenja. Korišteni su modeli eksponencijalnog zaglađivanja, ARIMA, Prophet i povratna neuronska mreža. Osim samih cijena, analizirani su i vanjski događaji koji su utjecali na promjene cijena kriptovaluta, a to je učinjeno analizom sentimenta objava s Twittera.

Abstract (english)

Cryptocurrencies are a digital form of currency managed by a decentralized market. The goal of many investors is to predict the prices of cryptocurrencies in advance and, using that, make a profit. Cryptocurrency prices can be viewed as a time series, and thus analyzed using time series analysis methods and tools. Time series analysis is a field of data mining that explores a series of data with the goal of finding interesting patterns in it and using them for other valuable tasks. The aim of this thesis is to predict the future prices of cryptocurrencies by analyzing time series using machine learning methods. Exponential smoothing, ARIMA, Prophet and recurrent neural networks were used. In addition to analysing cryptocurrency prices themselves, external events that influenced changes in the prices were also analyzed by using the sentiments of published Twitter posts.

• Keywords: kriptovalute; analiza vremenskih serija; strojno učenje
• Keywords (english): cryptocurrencies; time series analysis; machine learning
• Language: croatian
• URN:NBN: urn:nbn:hr:168:004593
• 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: 2023-07-13 10:20:52