K-Ants Clustering- A New Strategy Based on Ant Clustering

Abstract

In this article a new method of clustering with artificial ants is proposed. Unlike conventional ant clustering algorithms the true number of clusters must be provided for this algorithm in advance. Clustering is done using groups of ants which are as many as the number of clusters. The goal of each group is to collect members of one cluster. Two new functions are defined inspired from existing pick up and drop down probability functions which are used for inserting and removing loads of ants. Experimental results of the method demonstrated better accuracy in comparison to k-means and ant based clustering algorithm.