Enhance the Quality of Recommender Systems Using Collaborative Tagging

Abstract

newline Abstract newlineRecommender System (RS) is a recent emerging trend in Online Social Networks (OSN). RS is newlineclassified into Content-Based Filtering (CBF), in which the user s past behaviour is taken into newlineconsideration for the recommendation to the user for products. Collaborative Filtering (CF), in which newlinesimilar users or items are found to get better recommendations. Demographic Filtering (DMF), in newlinewhich demographic information is considered to find similar users for recommendation, Hybrid newlineFiltering, is the mixed technique of content based and collaborative filtering techniques. To avoid newlinecertain difficulty in CF, Tagging is introduced in which tags and tag cloud are special feature used for newlinerecommendations. Tags are special keywords used to describe the content in Online Social Networking newlinesites. Tag cloud is a set of tags which is used to represent most popular tags in the form of images or newlineshapes in recommendation websites. This tagging is additional information provided in CF which newlinehelps in finding similar users or similar items for recommendation with higher accuracy. The objective newlineof this thesis work is to address the problem of scalability, sparsity and cold start user/item issues newlinevarious similarity measures were applied to real datasets of Recommender Systems.