diplomsko delo
Timi Vovk (Author), Vili Podgorelec (Mentor)

Abstract

Strojno učenje se vse bolj uporablja za napoved prihodnjih stanj. V diplomskem delu je preizkušenih več modelov strojnega učenja za napoved vremenskih parametrov s ciljem izdelave čim boljšega. Modeli v enem mahu napovedujejo terminsko zračno temperaturo ali globalno sevanje za prvo, dvanajsto in štiriindvajseto uro. Ustvarjenih je več modelov z različno arhitekturo. Ti so naučeni iz obdelanih in urejenih podatkov pridobljenih iz javno dostopnega arhiva vremena ARSO. V ospredju algoritmov nadzorovanega stojnega učenja sta Elastic Net in GRU. Arhitektura za izdelavo GRU modelov je zgrajena na podlagi predlaganih modelov drugih avtorjev. Najbolje se je izkazal model GRU (60, 30, 10).

Keywords

nadzorovano strojno učenje;algoritem GRU;algoritem Elastic Net;temperatura zraka;sončno sevanje;diplomske naloge;

Data

Language: Slovenian
Year of publishing:
Typology: 2.11 - Undergraduate Thesis
Organization: UM FERI - Faculty of Electrical Engineering and Computer Science
Publisher: [T. Vovk]
UDC: 004.85.:519.216(043.2)
COBISS: 83605763 Link will open in a new window
Views: 232
Downloads: 28
Average score: 0 (0 votes)
Metadata: JSON JSON-RDF JSON-LD TURTLE N-TRIPLES XML RDFA MICRODATA DC-XML DC-RDF RDF

Other data

Secondary language: English
Secondary title: Weather forecast with the use of machine learning
Secondary abstract: Machine learning is often used for future state predictions. In this work there are several different models for weather prediction. Models predict air temperature and global irradiation. The models predict in one go air temperature or global irradiation for first, twelfth and twenty-fourth hour. Multiple models are created with different architecture. These are created from cleaned data from publicly accessible archive of weather data ARSO. Main algorithms used are GRU and elastic net. The GRU architecture is build from models that other authors suggested. The most successful one of them was GRU (60, 30, 10).
Secondary keywords: supervised machine learning;GRU;Elastic Net;air temperature;global irradiance;
Type (COBISS): Bachelor thesis/paper
Thesis comment: Univ. v Mariboru, Fak. za elektrotehniko, računalništvo in informatiko, Računalništvo in informacijske tehnologije
Pages: X, 44 str.
ID: 13294387