A Task Decomposition Algorithm Using Radial Basis Functions for Classification Problems

This paper proposes an algorithm for decomposing a multi-class classification problem into a set of two-class classification problems. The algorithm divides a set of input pattern vectors in each class into subsets according to the distribution of the selected input pattern vectors. The distribution is represented by RBF network, whose parameters are estimated according to the evaluation based on MDL criterion. In this paper, the algorithm applied for constructing a modular neural network. Experimental results showed that the algorithm simplifies multi-class classification problems effectively.