The identification of peers in a social network has been analyzed over giving importance to recognition in the way of connection and interaction between people. The coterie identification can be done by the structure of Facebook public pages. There are a lot of activities people perform which are diverse from each other. Since the different activities can’t be identified, we choose newsgroup pages and focus on extraction of one set of users. For this, we use Label Propagation Algorithm, Fast Unfolding and Spectral Clustering methods for identification and compare the results of each method to get a fair understanding of which method calculates the no.of user nodes effectively. These algorithms base the results upon the posts, likes and comments for each user activity. In order to get the effective values as outputs we eliminate the isolated and enhance the quality of the peer network across multiple pages with the use of subgraphs containing nodes that are more densely linked to each other than to the remaining nodes of the graph. The peer identification problem is tackled by the above mentioned algorithms.