Indian Journal of Science and Technology

Abstract

Background/Objectives: This paper presents a review on the usage of social tags that can be employed for recommending in social networks. It also focusses on pros and some algorithms as solutions to the problems. Recommender systems are the one used for providing information regarding the categories that belongs to the items which makes use of fuzzy clustering, in which the membership degree of the object is not available in the calculated categories. If calculated categories are available it will provide good quality recommendations to the user. Methods/Analysis: The method suggested in this work is to utilize fuzzy logic to classify the similar users from the clustered users based on their profiles and taste of viewing similar items. Fuzzy logic here plays a vital role in providing recommendation to the similar users once the classification is done. Findings: Hence, by surveying some papers on recommendation in social networking, we propose a solution by creating a dynamic system which would provision recommendations to the user to make the domain selection more effectively. The domains refers to books, movies, songs, which are purely dynamic. This methodology helps to display the most viewed item as the top most item and displays the item ratings in the webpage. The main approach used here is fuzzy logic with the detailed representation of features of objects and modeling user profiles. Applications/Improvements: This type of dynamic domain system will provide better results when compared to the conventional recommendation systems as it allows users to give comments on various products on different domains. This type of recommendation can be consumed by many social networking sites for gaining page rating and it can be further improved in future by combining the online dictionary to classify the terms used in user rating.
Keywords: Collaborative Filtering, Fuzzy Logic, Recommender Systems, Simmilarity, Tags, User Profiles