Modular Clustering by Radial Basis Function Network for Complexity Reduction in System Modeling

Abstract

Clustering is one of the dominant techniques of exploratory data analysis and data driven dynamic system modeling. In the context of model complexity, one of the powerful clustering methods is the method of orthogonal least squares (OLS) applied to radial basis functions (RBF) network. However, conventional way of utilization of OLS learning for a set of complex data is not desirable even though the learning process might be feasible for the amount of data at hand. This is due to the fact that some singular associations in the data can easily obscure many interrelations of interest among the data and also because of this, the rest of the associations for identification can heavily be limited. Therefore, as novel RBF clustering for system modeling, a set of time-series data is divided into several subsets as modules so that each subset is subjected to clustering separately. The dominant clusters in each subset are accumulated throughout the modular processing of total data set. The newly formed reduced data set which comprises the patterns from the clusters of the subsets, is subjected to final RBF network clustering by OLS for hierarchical cluster gradation from the subsets to identify the input-output model being searched for. In this way RBF clustering is accomplished substantially fast and at the same time effective reduction in complexity is obtained. The paper deals with the details of the novel RBF clustering.