A Novel Genetic Algorithm and Particle Swarm Optimization for Data Clustering

Clustering techniques suffer from fact that once they are merged or split, it cannot be undone or refined. Considering the stability of the Genetic Algorithm and the local searching capability of Swarm Optimization in clustering, these two algorithms are combined. Genetic Algorithms, being global search technique, have been widely applied for discovery of clusters. A novel data clustering based on a new optimization scheme which has benefits of high convergence rate and easy implementation method is been proposed were in local minima is disregarded in an intelligent manner. This paper, we intend to apply GA and swarm optimization (i.e., PSO) technique to optimize the clustering. We exemplify our proposed method on real data sets from UCI repository. From experimental results it can be ascertained that combined approach i.e., PSO_GA gives better clustering accuracy compare to PSO-based method.