Performance Analysis of Clustering Using Modified Grey Wolf Optimization

Abstract

Data clustering is the widely used technique in academia and industry to analyse large volumes of data with unknown patterns. Data clustering approaches that draw inspiration from biology are increasingly widely used. In this research, we propose a parallelized automated data clustering using a modified Grey Wolf Optimization technique based on the hunting style of the grey wolf which involves Tracking, chasing and approaching the prey. It will find the optimal solution from the generated ‘N’ solutions. However, the nature of massive data available in the repositories is unknown. So, it is a tedious task to guess the right number of clusters for the massive data. By repeating the procedure with clusters K = 2 to N, the suggested technique determines the best number of clusters. The ideal number of clusters are detected based on the best values of Silhouette index, Davies-Bouldin index and Calinski-Harabasz index. This research aims to propose a more efficient Intelligent clustering framework. The suggested approach operates in both the scenarios, i.e. with a predetermined number of clusters and an uncertain number of clusters. The user can either fix the number of clusters or let the system identify the optimal number of clusters. The proposed method parallelizes and automates cluster analysis in the most effective manner for determining the best clusters and forming natural clusters.