Szczegóły publikacji
Opis bibliograficzny
Neural system for power load prediction in a week time horizon / Andrzej BIELECKI, Marcin LENART // W: Artificial intelligence and soft computing : 15th international conference, ICAISC 2016 : Zakopane, Poland, June 12–16, 2016 : proceedings, Pt. 1 / eds. Leszek Rutkowski, [et al.]. — Switzerland : Springer International Publishing, cop. 2016. — (Lecture Notes in Artificial Intelligence ; ISSN 0302-9743 ; 9692). — ISBN: 978-3-319-39377-3; e-ISBN: 978-3-319-39378-0. — S. 25–34. — Bibliogr. s. 33–34, Abstr. — Toż na Dysku Flash
Autorzy (2)
Słowa kluczowe
Dane bibliometryczne
| ID BaDAP | 98350 |
|---|---|
| Data dodania do BaDAP | 2016-06-20 |
| DOI | 10.1007/978-3-319-39378-0_3 |
| Rok publikacji | 2016 |
| Typ publikacji | materiały konferencyjne (aut.) |
| Otwarty dostęp |  |
| Konferencja | International Conference on Artificial Intelligence and Soft Computing 2016 |
| Czasopismo/seria | Lecture Notes in Computer Science |
Abstract
In this paper a neural system for predicting electric power load in Poland in a week time horizon is presented. The system consists of seven multi-layer neural networks that have common input. Each network is dedicated to predict the total load in one of the seven successive days. Various form of input vectors as well as various ways of encoding them were tested. Verification which type of input data are crucial as well as which periodic aspects should be taken into account in data representation in week prediction was studied. Various numbers of neurons in a hidden layer were tested as well. The mean absolute percentage error (MAPE) is equal to 2.6% for the most effective system. © Springer International Publishing Switzerland 2016.
