Recently, a novelty multinomial logistic regression method wherethe initial covariate space is increased by adding the nonlinear transforma-tions of the input variables given by Gaussian Radial Basis Functions (RBFs)obtained by an Evolutionary Algorithm was proposed. However, there stillexist some problems with the standard Gaussian RBF, for example, the ap-proximation of constant valued functions or the approximation of high dimen-sionality associated to some real problems. In order to face of these problems,we propose the use of the Generalized Gaussian RBF (GRBF) instead ofthe standard Gaussian RBF. Our approach has been validated with a realproblem of disability classification, to evaluate its effectiveness. Experimen-tal results show that this approach is able to achieve good generalizationperformance.
展开▼
