Primerjava modelov za napovedovanje porabe električne energije

diplomsko delo

Nik Novak (Author), Damjan Strnad (Mentor), Štefan Kohek (Co-mentor)

Abstract

Uspešno napovedovanje porabe električne energije je pomembno z vidika ohranitve planeta, saj zaradi ustvarjanja viška porabljamo vire brez razloga. V diplomskem delu smo primerjali dva modela za napovedovanje porabe električne energije, in sicer nevronsko mrežo LSTM in model SARIMA za napovedovanje vrednosti v časovnih vrstah. Za testiranje modelov so bili uporabljeni podatki v tedenski ločljivosti, pridobljeni od podjetja Maked Energea, d. o. o. V rezultatih se je nevronska mreža LSTM pri uporabljenih nizih podatkov izkazala kot najboljša.

Keywords

strojno učenje;LSTM;RNN;SARIMA;napovedovanje;poraba električne energije;diplomske naloge;

Data

Language:	Slovenian
Year of publishing:	2019
Typology:	2.11 - Undergraduate Thesis
Organization:	UM FERI - Faculty of Electrical Engineering and Computer Science
Publisher:	[N. Novak]
UDC:	621.311.68(043.2)
COBISS:	22786582
Views:	715
Downloads:	87
Average score:	0 (0 votes)
Metadata:

Other data

Secondary language:	English
Secondary title:	Comparison of models for electricity consumption forecasting
Secondary abstract:	Electricity consumption forecast is important for conservation of Earth, as excess energy production is using the resources for no reason. In this thesis, we compared two models for predicting power consumption, namely the LSTM neural network and the SARIMA model for time series forecasting. Tests were performed on weekly resolution data obtained from Maked Energea, d.o.o. In the results, the LSTM neural network showed the best performance on the used datasets.
Secondary keywords:	machine learning;LSTM;RNN;SARIMA;forecasting;energy consumption.;
Type (COBISS):	Bachelor thesis/paper
Thesis comment:	Univ. v Mariboru, Fak. za elektrotehniko, računalništvo in informatiko, Računalništvo in informacijske tehnologije
Pages:	VI, 32 str.
ID:	11220710