Research on recommendation algorithm based on UserAttri-PCS

Aiming at two kinds of problems: cold start and data sparsity in the traditional collaborative filtering algorithm, this paper calculates similarity of attribute features and proportion feature as two recommendation basis to solve the above problems based on the user's personal information and their preference for item attribute. UserAttri-PCS algorithm in this paper is proposed based on the weight of two parts of the recommendation, and is revised according to the user's rating habits and criteria. Experimental results verify the effectiveness of the UserAttri-PCS recommendation algorithm. Compared with the traditional user-based collaborative filtering algorithm, UserAttri-PCS can improve the recommendation performance and its accuracy.