Proposing a new technique of group recommendation for movie data based on fuzzy rough clustering

Abstract

Group Recommendation is a technique in the Recommender Systems aiming at a group of users’ preferences. Via the study results of the concept of lower approximation and approximation of the fuzzy rough sets theory, the paper presents a Group Recommender Model that uses the nearest neighbor search algorithm based on the fuzzy rough sets to improve the recommender results. Also, the study is trying to perform clustering of items and select groups of itemsets that are of interest to a wide range of users. We have experimented on a specific movie data set and compared the results with the method based on searching traditional k-neighbors, thereby showing the effectiveness of our proposed model.