An Efficient Cluster Centroid Initialization Method for K-Means Clustering

Cluster analysis is one of the fundamental data analysis methods and K-Means is one of the most well-known popular clustering algorithms. The clustering result of the K-Means clustering algorithm is based on the correctness of the initial centroids, which are selected randomly. The original K-Means algorithm converges to local optimum, not the global optimum. The K-Means clustering performance can be enhanced if the initial cluster centers are found to it a series of procedure is done. Data in a cell is partitioned using a cutting plane that divides cell in two smaller cells. In this paper a new method is proposed for finding the better initial centroid and to estimate Number of Clusters based on two-cluster model which provides an efficient way of assigning the data points to suitable clusters with reduced time complexity. According to the experimental results, the proposed technique estimate the number of clusters and compute initial cluster centers for K-Means clustering. The achieved clustering results have more accuracy of clustering with less computational time when comparing to original K-Means clustering algorithm and CCIA method.