On L/sub 1/ convergence rate of RBF networks and kernel regression estimators with applications in classification

Studies convergence properties of radial basis function (RBF) nets in nonlinear estimation problems with parameters learned by minimizing the empirical mean integrated absolute error (MIAE). The authors show that MIAE of RBF nets converges to zero as the size of network and the size of the training sequence tends to infinity. The authors also provide the upper bound for the convergence rate for approximating smooth functions of order q. The obtained results are also applied in nonparametric classification.