K-harmonic Means Data Clustering with Particle Swarm Optimization

Unlike K-means, the K-Harmonic means (KHM) is less sensitive to initial conditions. However, KHM as a center-based clustering algorithm can only generate a local optimal solution. In this paper, we develop a new hybrid clustering algorithm combining Particle Swarm Optimization and K-Harmonic Means (HPSO) for solving this problem. This algorithm has been implemented and tested on several real datasets. The performance of this algorithm is compared with KHM and PSO. Our computational simulations reveal the HPSO clustering algorithm combines the ability of global searching of the PSO algorithm and the fast convergence and less sensitive to initial conditions of the KHM algorithm. The HPSO is a robust clustering algorithm.