Radial Basis Function Network in Reproducing Kernel Hilbert Space
Access this Article
Search this Article
The present study employs an idea of mapping data into a high dimensional feature space which is known as Reproducing Kernel Hilbert Space (RKHS), then performs Radial Basis Function (RBF) network in the feature space where the new basis function will be obtained and finally, Orthogonal Least Squares (OLS) method is employed to select a suitable set of centers (regressors) from a large set of candidates in order to obtain a sparse regression model in the feature space. The proposed method is employed to the simple scalar function estimation problems and nonlinear system identification problem by simulations.
- Research reports on information science and electrical engineering of Kyushu University
Research reports on information science and electrical engineering of Kyushu University 10(1), 9-14, 2005-03