An Improved Cohesion Based Community Recommendation System

Abstract

Social Networking Sites (SNS) are the dominating entities in the modern web. The importance of social networking sites in our life is increasing day by day as they are attracting millions of users by their interesting features and activities. It enables the researchers to use the information available in these sites. Online Community is appealing to people as they can enjoy sharing their ideas, view, and know about view of other people. At the same time, they are interested in joining different community. However, with the rapid growth of SNS’s resulting in information overload people are in dilemmas to choose right community from huge list of available communities and it is also time consuming. Potential choice of communities is influenced by many factors of user behavior and activeness in Social Networking Sites. The recent surge of research in recommendation algorithms is not surprising. But these algorithms have unsatisfactory results in community recommendation because of lack of intuition in judging rational behavior. Many researches are going on this point to find out recommendation system in various ways. To solve this problem, we introduce cohesion based community recommendation system. In this paper we design a general framework of community recommendation based on cohesion after analyzing the present methods of community recommendation. The main idea of the proposed approach is consisted of following stages- measuring friendship factor, measuring user factor, calculating threshold from present communities of user, community recommendation based on threshold, result analysis. We validated our idea on a small network in Facebook.