Tag-based Personalized Collaborative Movie Recommender System

A recommendation system is a set of a program used to help a user to discover products and contents which are relevant to their search query. Nowadays, the recommender system is accessible in every field of human life like E-Commerce, stock market, music, and in many more applications. Due to versatile uses of these applications lead to the acceleration of an enormous amount of data. Collaborative Filtering (a technique of recommendation system) (CF) has been broadly studied and applied to predict the interest of users by predicting the recommendation for the user. In this proposed approach, we introduce a hybrid recommendation system model which includes the methodology of collaborative and content-based filtering approach. The proposed method overcomes the well-known recommendation system problem i.e., sparsity, where users are not able to get an accurate recommendation, by recommending item to the user on basis of tag information. For the top-k recommendation generation, we proposed a tag value prediction approach which is used for topn recommendation generation according to their predicted tag value. Our approach performs better than the baseline approach with the F-1-score of 0.379 and the baseline F-1-score is 0.134.