On a maximizing model of spherical Bezdek-type possibilistic c-means and fuzzy multi-medoids clustering

In this study, two clustering frameworks are proposed based on a maximizing model of spherical Bezdek-type fuzzy clustering are proposed. One using possibilistic c-means, and the other using multi-medoids. In each framework, the basic model and its kernelization are presented, along with an appropriate spectral clustering technique. Kernelization allows the frameworks to capture nonlinear-bordered clusters, while spectral clustering solves their local convergence problems. Numerical examples demonstrate that the proposed frameworks produce good clustering results when an adequate parameter values are selected.