An Efficient JAYA-Based Clustering Technique

Abstract

This paper proposes a partitioned clustering technique using a relatively new evolutionary computational technique known as JAYA optimization algorithm. K-means clustering techniques happen to be widely researched partitional clustering algorithm in data mining literature. Despite its ease of implementation and widespread application, it has various shortcomings. The efficiency of k-means is heavily dependent on selection of initially cluster centroids. This apart, the time complexity of k-means is dependent on the size of the data sets to be clustered. In this work, we propose JAYA, a population-based approach which has scope to choose many candidate centroids and the initialization time and processed to evolve optimal cluster centroids and through simulation is carried out to benchmark data sets, and a comparison study is taken that JAYA-based algorithm is able to provide optimal performance, i.e., intra-cluster distance (ICD). The results are tabulated in the paper.