## Bayesian Radial Basis Functions of Unknown Dimension (1997)

Citations: | 3 - 1 self |

### BibTeX

@TECHREPORT{Holmes97bayesianradial,

author = {C.C. Holmes and B.K. Mallick},

title = {Bayesian Radial Basis Functions of Unknown Dimension},

institution = {},

year = {1997}

}

### OpenURL

### Abstract

A Bayesian framework for the analysis of radial basis functions (RBF) is proposed which readily accommodates uncertainty in the dimension of the model. A distribution is defined over the space of all RBF models of a given basis function and posteriors are computed using reversible jump Markov chain Monte Carlo samplers (Green, 1995). This alleviates the need to select one particular architecture during the modeling process. We show that the resulting models are relatively free from user set design parameters and that they exhibit good performance characteristics on a number of benchmark test series. Keywords: Radial basis functions, variable architecture selection, Bayesian neural networks, reversible jump Markov chain Monte Carlo. 1 Introduction We are interested in the regression of a target (response/output/dependent) variable y on an input (covariate/predictor/independent) set of variables x given the data set D of data pairings (y 1 ; x 1 ); (y 2 ; x 2 ) : : : (y N ; xN ). We...