A collaborative filtering recommendation method to the loyal-user problem

Although a variety of recommendation techniques have been proposed in recent years, collaborative filtering (CF) has been considered to be the most successful recommendation technique. Data sparsity is one of the problems which attract most attention from researchers. Data sparsity means there are lots of users having no or a small amount of rated item. Prediction to rating score of user´s unrated items is imprecision. In contrast to the data sparsity problem, what is called as loyal-user problem in this paper is neglected. In this paper, the loyal-user problem is analyzed and an approach is proposed to solve the problem. The experimental results also demonstrate the effectiveness of the method.