Diversity based self-adaptive clusters using PSO clustering for crime data

Diversity is the key parameter which plays the important role in defining the exploration capability of natural computing algorithms. Poor convergence is guaranteed, once diversity has lost prematurely. It is also true that there are number of sensitive parameters available with all paradigms of natural computing, whose optimal values drives the quality of solution. In this proposed work, diversity based self-adaption has been applied to particle swarm optimization to obtain better clusters. This diversity has been achieved with parameters like inertia weight, social and cognition constant. The proposed work has been applied over numeric benchmark and cluster data set to validate. Also new algorithm has been applied on crime datasets of Karnataka and Bengaluru to determine similar and different crime characteristics. © 2019, Bharati Vidyapeeth's Institute of Computer Applications and Management.