Redefining and Enhancing K-means Algorithm

Nimrat Kaur Sidhu, Rajneet kaur

This paper aims at finding the value of number of clusters in advance and to increase the overall performance of K-means algorithm. Although there are various methods for removing the disadvantages of k-means algorithm as the main problem is how to calculate the value of number of clusters in advance, secondly how to remove noisy data and outliers etc. There are many solutions also for eliminating these limitations as we have used various methods and validity indexes for but in this paper we are using the matrix method for removing these problems. Experiments show that this method provides very effective results.