Self-splitting competitive learning for RBF network and speech data clustering

The self-splitting competitive learning (SSCL) is a powerful algorithm that can be used in various problems, including unsupervised classification, curve detection and image segmentation. It solves the difficult problem of determining the number of clusters and the sensitivity to prototype initialization in clustering. In this paper, we apply SSCL to constructing a radial basis function (RBF) and speech data clustering. The experimental results show that SSCL performs well when used for training a RBF network and for speech data clustering.