A Hybrid Recommendation Method Combining User Profile and Product Profile

Abstract

In view of the cold start problem, the sparse rating and the deviation of emotions in the answers of the same rating of personalized product recommendation service in the e-commerce field, a personalized product recommendation method HRUP, which integrates user portrait and product portrait, was proposed. Firstly, user portrait was constructed from the basic attributes, interactive attributes, feedback attributes, interest attribute and situation attribute. Secondly, the product portrait was constructed based on five aspects: the user characteristics of the product, the popularity of the product, the feedback effect of the product, and the purchase situation of the product. Then, the candidate recommended product set was obtained by user similarity calculation. Next, the product similarity was used to sort and recommend the products. The experiment results show that the MAE and RMSE values of the proposed method are lower than those of the collaborative filtering method and the recommendation model based on single user profile or product profile.