Development of design strategy for rbf neural network with the aid of context-based FCM

Abstract

In this paper, we develop a new design strategy of Radial Basis Function (RBF) neural network and provide a comprehensive design methodology and algorithmic setup supporting its development. The architecture of the network is fully reflective of the structure encountered in the training data which are granulated with the aid of clustering techniques. More specifically, the output space is granulated with use of FCM clustering while the information granules in the multidimensional input space are formed by using a so-called context-based Fuzzy C-Means which takes into account the structure being already formed in the output space. A series of numeric studies exploiting synthetic data and data from the Machine Learning Repository provide a detailed insight into the nature of the algorithm and its parameters as well as offer some comparative analysis. © 2009 Springer Berlin Heidelberg.