Combining Multiple Criteria and Multidimension for Movie Recommender System
Keittima Chapphannarungsri, Saranya Maneeroj · 2009
Most of current Recommender Systems based on Content-Based Filtering, Collaborative Fil- tering, Demographic Filtering and Hybrid Filtering which are concentrated on user and item entities. Many research papers are improved by pointing out either Multiple Criteria Rating approach or Multidi- mensional approach for Recommender System. This paper proposes an advanced Recommender System to provide higher quality of recommendations by com- bining the Multiple Criteria rating and the Multidi- mensional approaches. For the Multiple Criteria ap- proach, this paper proposed a method that changes the way of weighting to be more suitable and also con- cern about the frequency of the selection movie fea- tures. To do Multidimensional approach, the Multi- ple Linear Regression is applied to analyze the contex- tual information of user characteristics. According to the experimental evaluation, the combining of Multi- ple Criteria Rating and Multidimensional approaches provide more accurate recommendation results than the current Hybrid Recommender Systems.