Uporaba metoda strojnog učenja nad podacima vremenskih serija temperatura

master's thesis

• Title (english): Using machine learning methods on temperature time series data
• Author: Ivan Marović
• Mentor: Boris Vrdoljak (mentor)
• Committee member: Boris Vrdoljak (predsjednik povjerenstva)
• Committee member: Mirta Baranović (član povjerenstva)
• Committee member: Ljiljana Brkić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2022-07-01, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

U radu se obrađuje korištenje metoda dubokog učenja kako bi se poboljšalo predviđanje temperature nekolicine meteoroloških web servisa temeljem povijesnih podataka tih web servisa. Podaci koji su korišteni u radu su vremenske serije, tako da se na početku rada govori o tome što su vremenske serije i na koje načine ih je moguće analizirati. Potom su opisane metode strojnog i dubokog učenja te matematička pozadina istih. Nakon toga su dodatno objašnjene podvrste neuronske mreže koje se mogu koristiti u dubokom učenju zvane povratne neuronske mreže, među kojima se nalazi i LSTM – poboljšanje nad radom običnih povratnih neuronskih mreža. Zatim je objašnjeno programsko rješenje modela te su prezentirani rezultati, da bi na kraju još bila prikazana korisnička aplikacija koja koristi napravljene modele.

Abstract (english)

This thesis examines how deep learning methods can be used for the improvement of temperature predictions of a few meteorological web services based on historic data of those web services. Time series data were used in model creation so the thesis first covers what time series are and in which way could they be analysed. Next, machine and deep learning methods are discussed. After that, the thesis focuses on recurrent neural networks among which is LSTM – an improvement over regular recurrent neural networks. Following that, the software solution is explained and the results are presented. Finally, a user application that uses models created in the software solution is presented.

• Keywords: strojno učenje; duboko učenje; povratne neuronske mreže; LSTM; temperatura; vremenske serije
• Keywords (english): machine learning; deep learning; recurrent neural networks; LSTM; temperature; time series
• Language: croatian
• URN:NBN: urn:nbn:hr:168:574296
• 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 13:01:17