International forum of researchers Students and Academician
In this paper, the authors reflect the comparative study of k-means clustering algorithm with the mean based initial centroids. The original k-means algorithm helps to calculate the distance between the data objects, but the difficulty is that the k-means algorithm is not efficient and accurate to calculate distance and henceforth make clusters. K-means algorithm needs lots of iterations to make clusters. In this paper, centroids based algorithm is used to avoid lots of iterations. This systematic method will help to give the global optimal result for any set of data objects. The limitation of k-means algorithm will be removed by using initial centroids.