Entrenamiento supervisado de redes neuronales de impulsos
- Lucas, Sergio 1
- Portillo, Eva 1
- Zubizarreta, Asier 1
- Cabanes, Itziar 1
-
1
Universidad del País Vasco/Euskal Herriko Unibertsitatea
info
Universidad del País Vasco/Euskal Herriko Unibertsitatea
Lejona, España
- Carlos Balaguer Bernaldo de Quirós (coord.)
- José Manuel Andújar Márquez (coord.)
- Ramon Costa Castelló (coord.)
- Carlos Ocampo Martínez (coord.)
- Jesús Fernández Lozano (coord.)
- Matilde Santos Peñas (coord.)
- José Enrique Simó Ten (coord.)
- Montserrat Gil Martínez (coord.)
- Jose Luis Calvo Rolle (coord.)
- Raúl Marín Prades (coord.)
- Eduardo Rocón de Lima (coord.)
- Elisabet Estévez Estévez (coord.)
- Pedro Jesús Cabrera Santana (coord.)
- David Muñoz de la Peña Sequedo (coord.)
- José Luis Guzmán Sánchez (coord.)
- José Luis Pitarch Pérez (coord.)
- Oscar Reinoso García (coord.)
- Oscar Déniz Suárez (coord.)
- Emilio Jiménez Macías (coord.)
- Vanesa Loureiro Vázquez (coord.)
Publisher: Servizo de Publicacións ; Universidade da Coruña
ISBN: 978-84-9749-841-8
Year of publication: 2022
Pages: 216-223
Congress: Jornadas de Automática (43. 2022. Logroño)
Type: Conference paper
Abstract
This article proposes a new supervised training strategy with Spiking Neural Networks (SNN) for time series forecasting. Currently, the vast majority of work on SNN is focused on classification tasks, especially image classification. In this sense, this paper is one of the first works to apply SNN for time series forecasting, whose results are very promising. Two benchmark datasets have been used to validate the methodology: IBM stock market information and EEG signals. Among the results, it is shown that the performance of SNN depends, as expected, on the dynamic of the signal or time series to be forecast.