A New Kernelized Fuzzy C-Means Clustering Algorithm with Enhanced Performance

Recently Kernelized Fuzzy C-Means clustering technique where a kernel-induced distance function is used as a similarity measure instead of a Euclidean distance which is used in the conventional Fuzzy C-Means clustering technique, has earned popularity among research community. Like the conventional Fuzzy C-Means clustering technique this technique also suffers from inconsistency in its performance due to the fact that here also the initial centroids are obtained based on the randomly initialized membership values of the objects. Our present work proposes a modified method to remove the effect of random initialization from Kernelized Fuzzy C-Means clustering technique and to improve the overall performance of it. In our proposed method we have used the algorithm of Yuan et al. to determine the initial centroids. These initial centroids are then used in the conventional Kernelized Fuzzy C-Means clustering technique to obtain the final clusters. We have also provided a comparison of our method with the Kernelized Fuzzy C-Means clustering technique of Hogo using two validity measures namely Partition Coefficient and Clustering Entropy.