Časovne vrste in razvrščanje z zavrnitvijo

Kovačič, Nace (author), Zupan, Blaž (mentor)

V diplomskem delu predstavimo pristop k napovedovanju dogodkov drastičnih sprememb valutnega tečaja kriptovalut. Za napovedovanje uporabimo modela XGBClassifier in konvolucijsko nevronsko mrežo. Primerjamo njuno točnost napovedi in spremembo točnosti napovedi pri napovedovanju z možnostjo zavrnitve.

• Language: Slovenian
• Keywords: časovne vrste, časovne vrste z zavrnitvijo, konvolucijske nevronske mreže
• Work type: Bachelor thesis/paper
• Typology: 2.11 - Undergraduate Thesis
• Organization: FRI - Faculty of Computer and Information Science
• Year: 2023
• PID: 20.500.12556/RUL-152731
• COBISS.SI-ID: 169130755
• Publication date in RUL: 04.12.2023

Secondary language

• Language: English
• Title: Time Series with Classification and Rejection

Abstract

In this thesis we present an approach to predicting events of drastic changes in the exchange rate of cryptocurrencies. For prediction we use the XGBClassifier model and the convolutional neural network. We compare their prediction accuracy and the change in prediction accuracy when predicting with the option of rejection.

• Keywords: timeseries, timeseries with rejection, convolutional neural networks