Podržano učenje u trgovanju imovinom zasnovano na dubokom modelu s latentnim varijablama

Kaštelan, Niko

prikaz prve stranice dokumenta Podržano učenje u trgovanju imovinom zasnovano na dubokom modelu s latentnim varijablama

No public access

master's thesis

Podržano učenje u trgovanju imovinom zasnovano na dubokom modelu s latentnim varijablama

Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing, 2024. urn:nbn:hr:168:022001

Kaštelan, Niko

University of Zagreb
Faculty of Electrical Engineering and Computing

Cite this document

APA 6th Edition

Kaštelan, N. (2024). Podržano učenje u trgovanju imovinom zasnovano na dubokom modelu s latentnim varijablama (Master's thesis). Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing. Retrieved from https://urn.nsk.hr/urn:nbn:hr:168:022001

MLA 8th Edition

Kaštelan, Niko. "Podržano učenje u trgovanju imovinom zasnovano na dubokom modelu s latentnim varijablama." Master's thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, 2024. https://urn.nsk.hr/urn:nbn:hr:168:022001

Chicago 17th Edition

Kaštelan, Niko. "Podržano učenje u trgovanju imovinom zasnovano na dubokom modelu s latentnim varijablama." Master's thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, 2024. https://urn.nsk.hr/urn:nbn:hr:168:022001

Harvard

Kaštelan, N. (2024). 'Podržano učenje u trgovanju imovinom zasnovano na dubokom modelu s latentnim varijablama', Master's thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, accessed 07 February 2025, https://urn.nsk.hr/urn:nbn:hr:168:022001

Vancouver

Kaštelan N. Podržano učenje u trgovanju imovinom zasnovano na dubokom modelu s latentnim varijablama [Master's thesis]. Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing; 2024 [cited 2025 February 07] Available at: https://urn.nsk.hr/urn:nbn:hr:168:022001

IEEE

N. Kaštelan, "Podržano učenje u trgovanju imovinom zasnovano na dubokom modelu s latentnim varijablama", Master's thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, Zagreb, 2024. Available at: https://urn.nsk.hr/urn:nbn:hr:168:022001

Cite this item: https://urn.nsk.hr/urn:nbn:hr:168:022001

Please login to the repository to save this object to your list.

Metadata

Title	Podržano učenje u trgovanju imovinom zasnovano na dubokom modelu s latentnim varijablama
Title (english)	Reinforcement learning in asset trading based on deep latent variable model
Author	Niko Kaštelan
Mentor	Zvonko Kostanjčar (mentor)
Committee member	Zvonko Kostanjčar (predsjednik povjerenstva)
Committee member	Tomislav Petković (član povjerenstva)
Committee member	Stjepan Begušić (član povjerenstva)
Granter	University of Zagreb Faculty of Electrical Engineering and Computing Zagreb
Defense date and country	2024-07-04, Croatia
Abstract	Metode dubokog podržanog učenja ekstenzivno se koriste u rješavanju problema optimalnog trgovanja imovinom. Duboki modeli s latentnim varijablama pokazali su se vrlo uspješnima u modeliranju pozadinskih generativnih procesa kompleksnih skupova podataka. U ovom radu predstavljen je novi pristup modeliranju distribucija budućih povrata imovine na temelju vrijednosti iz prošlog razdoblja korištenjem dubokog modela s latentnim varijablama. Potom su korištenjem algoritma dubokog podržanog učenja TD3 na temelju modeliranih distribucija naučene strategije trgovanja imovinom. Pokazano je da naučene strategije trgovanja imovinom svojim performansama ne nadmašuju performanse samog tržišta u značajnoj mjeri.
Abstract (english)	Deep reinforcement learning methods have been extensively used in solving the problem of optimal asset trading. Deep latent variable models have proven very successful in modeling the underlying generative processes of complex datasets. This work presents a novel approach to modeling the distributions of future asset returns that is based on the values observed in the past using a deep latent variable model. The modeled distributions were then used to learn asset trading strategies using the deep reinforcement learning algorithm TD3. It was shown that the performance of the learned asset trading strategies does not exceed that of the market itself by a meaningful margin.
Keywords
Keywords (english)
Language	croatian
URN:NBN	urn:nbn:hr:168:022001
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
Terms of use
Public note
Repository	FER Repository
Created on	2024-10-12 21:01:51