Binary flower pollination algorithm and its application to feature selection
Carregando...
Data
2015-01-01
Orientador
Coorientador
Pós-graduação
Curso de graduação
Título da Revista
ISSN da Revista
Título de Volume
Editor
Tipo
Artigo
Direito de acesso
Acesso aberto
Resumo
The problem of feature selection has been paramount in the last years, since it can be as important as the classification step itself. The main goal of feature selection is to find out the subset of features that optimize some fitness function, often in terms of a classifier’s accuracy or even the computational burden for extracting each feature. Therefore, the approaches to feature selection can be modeled as optimization tasks. In this chapter, we evaluate a binary-constrained version of the Flower Pollination Algorithm (FPA) for feature selection, in which the search space is a boolean lattice where each possible solution, or a string of bits, denotes whether a feature will be used to compose the final set. Numerical experiments over some public and private datasets have been carried out and comparison with Particle Swarm Optimization, Harmony Search and Firefly Algorithm has demonstrated the suitability of the FPA for feature selection.
Descrição
Palavras-chave
Idioma
Inglês
Como citar
Studies in Computational Intelligence, v. 585, p. 85-100.