Gutiérrez P.A. , Hervás C. , CARBONERO RUZ, MARIANO, Fernández J.C.
No
Neurocomputing
Article
Científica
1.44
0.479
01/01/2009
000268733700002
2-s2.0-71649107625
This paper proposes a hybrid neural network model using a possible combination of different transfer projection functions (sigmoidal unit, SU, product unit, PU) and kernel functions (radial basis function, RBF) in the hidden layer of a feed-forward neural network. An evolutionary algorithm is adapted to this model and applied for learning the architecture, weights and node typology. Three different combined basis function models are proposed with all the different pairs that can be obtained with SU, PU and RBF nodes: product-sigmoidal unit (PSU) neural networks, product-radial basis function (PRBF) neural networks, and sigmoidal-radial basis function (SRBF) neural networks; and these are compared to the corresponding pure models: product unit neural network (PUNN), multilayer perceptron (MLP) and the RBF neural network. The proposals are tested using ten benchmark classification problems from well known machine learning problems. Combined functions using projection and kernel functions are found to be better than pure basis functions for the task of classification in several datasets. © 2009 Elsevier B.V.
Classification (of information); Functions; Learning systems; Multilayer neural networks; Neural networks; Benchmark classification; Evolutionary neural network; Hybrid neural networks; Kernel basis functions; Machine learning problem; Multi layer perceptron; Product unit neural network; Projection basis functions; Radial basis function networks; article; artificial neural network; classification; data base; evolutionary algorithm; intermethod comparison; kernel method; machine learning; mathematical analysis; mathematical computing; molecular evolution; priority journal; statistical model