Por favor, use este identificador para citar o enlazar este ítem:
https://doi.org/10.1109/TNNLS.2023.3234629
Twittear
Título: | Surrogate-assisted and filter-based multi-objective evolutionary feature selection for deep learning |
Fecha de publicación: | 12-ene-2023 |
Editorial: | Institute of Electrical and Electronics Engineers |
Cita bibliográfica: | IEEE Transactions on Neural Networks and Learning Systems |
ISSN: | 2162-237X 2162-2388 |
Materias relacionadas: | CDU::6 - Ciencias aplicadas::68 - Industrias, oficios y comercio de artículos acabados. Tecnología cibernética y automática |
Palabras clave: | Feature selection deep learning surrogate- assisted multi-objective evolutionary algorithms time series forecasting air quality indoor temperature |
Resumen: | Feature selection for deep learning prediction mod- els is a difficult topic for researchers to tackle. Most of the ap- proaches proposed in the literature consist of embedded methods through the use of hidden layers added to the neural network architecture that modify the weights of the units associated with each input attribute so that the worst attributes have less weight in the learning process. Other approaches used for deep learning are filter methods, which are independent of the learning algorithm, which can limit the precision of the prediction model. Wrapper methods are impractical with deep learning due to their high computational cost. In this paper, we propose new attribute subset evaluation feature selection methods for deep learning of the wrapper, filter and wrapper-filter hybrid types, where multi-objective and many-objective evolutionary algorithms are used as search strategies. A novel surrogate-assisted approach is used to reduce the high computational cost of the wrapper-type objective function, while the filter-type objective functions are based on correlation and an adaptation of the reliefF algorithm. The proposed techniques have been applied in a time series forecasting problem of air quality in the Spanish south-east and an indoor temperature forecasting problem in a domotic house, with promising results compared to other feature se |
Autor/es principal/es: | Espinosa Fernández, Raquel Jiménez Barrionuevo, Fernando Palma Méndez, José Tomás |
Facultad/Departamentos/Servicios: | Facultades, Departamentos, Servicios y Escuelas::Departamentos de la UMU::Ingeniería de la Información y las Comunicaciones Facultades, Departamentos, Servicios y Escuelas::Facultades de la UMU::Facultad de Informática |
Versión del editor: | https://ieeexplore.ieee.org/document/10016286 |
URI: | http://hdl.handle.net/10201/127323 |
DOI: | https://doi.org/10.1109/TNNLS.2023.3234629 |
Tipo de documento: | info:eu-repo/semantics/article |
Número páginas / Extensión: | 15 |
Derechos: | info:eu-repo/semantics/openAccess |
Descripción: | © 2023, Publishers. This manuscript version is made available under the CC-BY-NC-ND 4.0 license http://creativecommons.org/licenses/by-nc-nd/4.0/ This document is the Accepted version of a Published Work that appeared in final form in IEEE Transactions on Neural Networks and Learning Systems. To access the final edited and published work see https://doi.org/10.1109/TNNLS.2023.3234629 |
Aparece en las colecciones: | Artículos: Ingeniería de la Información y las Comunicaciones |
Ficheros en este ítem:
Fichero | Descripción | Tamaño | Formato | |
---|---|---|---|---|
EspinosaSurrogateTNNLS_2023.pdf | Artículo publicado en IEEE Transactions on Neural Networks and Learning Systems | 1,06 MB | Adobe PDF | Visualizar/Abrir |
Supplementary materials.pdf | Suplementary materials | 696,08 kB | Adobe PDF | Visualizar/Abrir |
Este ítem está sujeto a una licencia Creative Commons Licencia Creative Commons