Movie recommender system using pseudo rating and multidimensional data

Abstract

This paper utilizes the Multi criteria Pseudo rating and Multidimensional user profile to enhance the quality and the accuracy of the recommender system. Recommender systems are usually classified into three categories based on how recommendations are made (i) Content – Based recommendations, (ii) Collaborative Filtering recommendations and (iii) Hybrid recommendations. To reduce the Sparsity Rating problem and fulfill the co-rated items in CF table, the current systems create the Pseudo ratings usually based on one criteria. This paper proposes pseudo ratings based on Multi criteria and also concentrates on the Contextual Information as Multidimensional. To do the Pseudo ratings based on Multi criteria, the Naïve Bayes is applied to classify the Multi criteria of user’s preference. To incorporate Multidimensional, the Multi regression is applied to analyze the contextual information of user. According to the experimental evaluation, the recommender system on movie domain called ModernizeMovie is created and shows that the Multi criteria Pseudo ratings and Multidimensional user profile enhance the quality and accuracy of recommendation results.