Hybrid Clustering Optimization of K-Means, Particle Swarm Optimization and Cuttlefish Optimization

Abstract

Predominant conventional clustering methods hang on to the distance metrics and similarity functions have the common issue to fit into the local optima and also their performance is based on the initial cluster centers. In this paper we present a new hybrid approach using K-Means, Particle Swarm optimization (PSO) and Cuttlefish Optimization (CFA) to enhance the clustering process. The data sets are taken from the UCI repository for experimentation. To validate the performance of the proposed approach it is compared based on the global fitness, Mean and Standard deviation (SD) with the well-known methods PSO, CFA and Hybrid-PSO. The proposed approach produces an impressive results over the existing methods.