Analytical Review of K-Means based Algorithms and Evaluation Methods

The traditional k-Means clustering algorithm suffers from many limitations and various researchers have proposed many revisions to improve its performance. In this paper we describe the k-Means clustering algorithm and explain its major limitations. We further study the revisions introduced by the researchers to the k-Means algorithm and explore in detail the parameters that are used to evaluate the performance of k-Means based clustering algorithms through experiments on several data sets.