Interest based Recommender System for Social Media

Abstract

In great popularity of Social Media, flourishing massive world of online social communities entrust users to formulate themselves over set of the interests.Users love to know about the people they like and the different things they are interested into.Thus, they require most appropriate source of information sharing and alert gathering.Recommender system is a solution for it.Every site does have a recommender system but these recommender systems are mostly based on the subscription choices of user"s friends.When a user searches for a particular interest over these sites, the first few results that come up to the screen are the broader interest groups which become over crowded over the time and the user remains with the choice to join an interest group which is of lesser interest than the appropriate one.Hence, this paper proposes a interest based Recommender System named as "MAAC" Recommender System, which recommends the users to follow the more specific interest feeds, solely based on the current subscribed interests of a user instead of demographic knowledge or Friends knowledge.Also, it recommends the interests which are closely related to the interests already linked with the user"s profile.The proposed recommender system proves to be highly effective in recommending the various interests to the user, thus incorporating the high percent accuracy in terms of recall, precision and F1 measure as compared to other recommender system.