Integrating User Behavior Model and Collaborative Filtering Methods in Recommender Systems

초록

The user predicting preference method using a collaborative filtering(CF) doesn't only reflect any contents about item but also solve the problem of the Sparsity and First-Rater. In this paper, we suggest the method for predicting using Bayesian estimated value and associative user clustering for the recalculation of preference to complement problems of the existing CF method. We suggest Representative Attributes - Neighborhood (RA-Neighborhood) method using for prediction finding the similar neighborhood through extracting the representative attributes that most affects the preference. We improved the efficiency by using the associative user clusting analysis for recalculating the preference of specific item within the cluster item vector to the CF algorithm. Associative users are clustered according to the genre through using ARHP algorithm, and new users are classified into one of these genre by Naive Bayes learning classifier. Besides, to get the similarity between users belonged to the classified genre and new users, this paper allows the different estimated value to the item which users evaluated through Naive Bayes learning. We evaluate our method on a large CF database of user rating and it significantly outperforms previous proposed method.