In this paper, we present a new algorithm to identify non- overlapping like-minded communities in a social network and compare its performance with Girvan-Newman algorithm, Lovain method and some well-known hierarchical clustering algorithms on Twitter and Filmtipset datasets.