ABK-means: an algorithm for data clustering using ABC and K-means algorithm

Abstract

Data clustering has become one of the promising areas in data mining field. The algorithms, such as K-means and FCM are traditionally used for clustering purpose. Recently, most of the research studies have concentrated on optimisation of clustering process using different optimisation methods. The commonly used optimising algorithms such as particle swarm optimisation and genetic algorithms have given some significant contributions for optimising the clustering results. In this paper, we have proposed an approach to optimise the clustering process using artificial bee colony ABC algorithm with K-means operator. Here, we modify the traditional ABC algorithm with K-means operator. From the experimental results, we conclude that our proposed approach has upper hand over other methods. The comparative analysis of our approach with other algorithms using datasets such as iris, thyroid and wine is satisfactory. The proposed approach has achieved the intra-cluster distance values of 68.2, 9,682.4 and 12,234.4 for iris, thyroid and wine datasets respectively for the best cases.