A new Kernelized hybrid c-mean clustering model with optimized parameters

A possibilistic approach was initially proposed for c-means clustering. Although the possibilistic approach is sound, this algorithm tends to find identical clusters. To overcome this shortcoming, a possibilistic Fuzzy c-means algorithm (PFCM) was proposed which produced memberships and possibilities simultaneously, along with the cluster centers. PFCM addresses the noise sensitivity defect of Fuzzy c-means (FCM) and overcomes the coincident cluster problem of possibilistic c-means (PCM). Here we propose a new model called Kernel-based hybrid c-means clustering (KPFCM) where PFCM is extended by adopting a Kernel induced metric in the data space to replace the original Euclidean norm metric. Use of Kernel function makes it possible to cluster data that is linearly non-separable in the original space into homogeneous groups in the transformed high dimensional space. From our experiments, we found that different Kernels with different Kernel widths lead to different clustering results. Thus a key point is to choose an appropriate Kernel width. We have also proposed a simple approach to determine the appropriate values for the Kernel width. The performance of the proposed method has been extensively compared with a few state of the art clustering techniques over a test suit of several artificial and real life data sets. Based on computer simulations, we have shown that our model gives better results than the previous models.