prikaz prve stranice dokumenta Primjena modela s dugom kratkoročnom memorijom na predviđanje performansi u višezadaćnom radu
No public access
undergraduate thesis
Primjena modela s dugom kratkoročnom memorijom na predviđanje performansi u višezadaćnom radu
Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing, 2024. urn:nbn:hr:168:362800
Panđa, Luka
University of Zagreb
Faculty of Electrical Engineering and Computing

Institutional repository: FER Repository

Cite this document

Panđa, L. (2024). Primjena modela s dugom kratkoročnom memorijom na predviđanje performansi u višezadaćnom radu (Undergraduate thesis). Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing. Retrieved from https://urn.nsk.hr/urn:nbn:hr:168:362800

Panđa, Luka. "Primjena modela s dugom kratkoročnom memorijom na predviđanje performansi u višezadaćnom radu." Undergraduate thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, 2024. https://urn.nsk.hr/urn:nbn:hr:168:362800

Panđa, Luka. "Primjena modela s dugom kratkoročnom memorijom na predviđanje performansi u višezadaćnom radu." Undergraduate thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, 2024. https://urn.nsk.hr/urn:nbn:hr:168:362800

Panđa, L. (2024). 'Primjena modela s dugom kratkoročnom memorijom na predviđanje performansi u višezadaćnom radu', Undergraduate thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, accessed 29 December 2024, https://urn.nsk.hr/urn:nbn:hr:168:362800

Panđa L. Primjena modela s dugom kratkoročnom memorijom na predviđanje performansi u višezadaćnom radu [Undergraduate thesis]. Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing; 2024 [cited 2024 December 29] Available at: https://urn.nsk.hr/urn:nbn:hr:168:362800

L. Panđa, "Primjena modela s dugom kratkoročnom memorijom na predviđanje performansi u višezadaćnom radu", Undergraduate thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, Zagreb, 2024. Available at: https://urn.nsk.hr/urn:nbn:hr:168:362800

Metadata
TitlePrimjena modela s dugom kratkoročnom memorijom na predviđanje performansi u višezadaćnom radu
Title (english)Application of models with long short-term memory to performance prediction in multitasking
AuthorLuka Panđa
MentorSiniša Popović (mentor)
Committee memberSiniša Popović (predsjednik povjerenstva)
Committee memberHrvoje Nuić (član povjerenstva)
Committee memberMate Gambiraža (član povjerenstva)
GranterUniversity of Zagreb
Faculty of Electrical Engineering and Computing
Zagreb
Defense date and country2024-09-02, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Computing
Abstract
Ovaj rad se fokusira na donošenje predviđanja kroz različite vremenske horizonte koristeći različite modele s dugom kratkoročnom memorijom. Koristeći PyTorch biblioteku unutar programskog jezika Python, konstruiramo različite arhitekture modela temeljenih na dugoj kratkoročnoj memoriji, provodimo njihovu evaluaciju koristeći standardne metrike evaluacije te predstavljamo metodu reprezentiranja diskretnih događaja u kontekstu vremenskih sekvenci. Modeli se pokazuju ograničeno uspješnima u isčekivanju događaja te donošenju predviđanja nad kraćim vremenskim horizontima. Također predlažemo potencijalna poboljšanja uz koja se modeli mogu prilagoditi na dulje vremenske horizonte te provoditi predviđanja s obzirom na različite ispitanike. Kroz budućnost ovaj rad može služiti kao temelj za donošenje predviđanja performansi korisnika te korisnikovih budućih akcija u pravom vremenu.
Abstract (english)
In this thesis we focus on time series prediction over different time horizons using multiple long short-term memory based models. Using the PyTorch library in Python, we construct different model architectures based on long short-term memory, conduct their evaluation using standard evaluation metrics and present a method for representing discrete events in a time series context. The models have shown limited success in anticipating events and conducting predictions over shorter time horizons. We also present potential improvements with which the models can better adapt to longer time horizons and make predictions on multiple examinees. In the future this thesis can be used as a basis for real time user performance prediction and the prediction of user's future actions.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:168:362800
Study programmeTitle: 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 resourceText
File originBorn digital
Access conditionsClosed access
Terms of useLicence In copyright icon
Public note
Created on2024-10-12 21:45:01