CFPRS: Collaborative Filtering Privacy Recommender System for Online Social Networks

Social-networking sites (SNSs) are known to be among the most prevalent methods of online communication. Owing to their increasing popularity, online privacy has become a critical issue for these sites. The tools presently being utilized for privacy settings are too ambiguous for ordinary users to understand and the specified policies are too complicated. In this paper, a collaborative filtering privacy recommender system is proposed. The implementation of the system was initiated by examining the users ’ attitudes toward privacy; whereby the most significant factors impacting users ’ attitudes towards privacy were determined to be location, religion and gender. The next step involved the classification of the users into various groups on the basis of the above factors. The paper presents a method of integrating the identified factors into the collaborative filtering algorithm to improve the filtering process. The evaluation of results reflects the accuracy of recommendations and proves that the use of the clustering model assisted the CF recommender in its creation of appropriate recommendations for each user