Hybrid Symbiotic Organism Search algorithms for Automatic Data Clustering

Abstract

Cluster analysis is an essential tool in data mining. Several clustering algorithms have been proposed and implemented for which most are able to find the good quality or optimal clustering solutions. However, most of these algorithms still depend on the number of a cluster being provided a priori. In dealing with real-life problems, the number of clusters is unknown and determining the optimal number of clusters for a large density and high dimensionality dataset is quite a difficult task to handle. This paper, therefore, proposes five new hybrid symbiotic organism search algorithms to automatically partition datasets without any prior information regarding the number of clusters. Furthermore, the hybrid algorithms will be evaluated in terms of solution quality using the Davies-Bouldin clustering validity index. The simulation results show that the performance of the hybrid symbiotic organisms search particle swarm optimization algorithm is superior to the other proposed hybrid algorithms.