Abstract: Data clustering is the process of grouping of data which are close together. The most popular clustering algorithm used in various domains is K-means. However, K-means algorithm has four main drawbacks: it converges to the local optimum solutions. The results obtained are strongly depends upon the selection of initial seeds, number of clusters need to be known in advance and it does not provide approximation guarantee. Various initialization methods were proposed to improve the performance of K-means algorithm. As the convergence of data points are only based on the selection of initial centroids, researchers are proposing an efficient algorithm for finding the initial centroids by considering distance on extreme ends, called K-means Minimum-Average-Maximum (K-MAM) Method. The proposed algorithm is tested with some of the UCI repository datasets and are compared with K-means and K-means++ algorithms. The results show that the proposed algorithm converges very fast with better accuracy.
S. Dhanabal and S. Chandramathi, 2013. An Efficient K-Means Initialization Using Minimum-Average-Maximum (MAM) Method. Asian Journal of Information Technology, 12: 77-82.