Personalized comment recommendation method based on user preferences

Abstract

The invention discloses a personalized comment recommendation method based on user preferences, and the method comprises the steps: 1, obtaining user information: providing personalized questions and answers for a new user who logs in for the first time, and obtaining the information of an old user, especially a product which the user is interested in, and concerned game content; and 2, carrying out product screening: according to the information of the user, recommending products which the user may be interested in by using a collaborative filtering algorithm; and step 3, carrying out comment recommendation: for product comments which the user is interested in, predicting the score of the comments by the user by adopting a personalized comment recommendation method based on singular value decomposition, recommending comments with high scores to the user, providing personalized questions and answers for a new user who logs in for the first time, and collecting the basic information, interest points and focus points of the old user. The method overcomes the defect that a traditional method is lack of individuation, can display interested and valuable high-quality comments to users, and achieves the effects of attracting the users and helping potential consumers to make purchase decisions.

Description

technical field [0001] The invention relates to the related technical field of network comment screening and push, in particular to a personalized comment recommendation method based on user preference. Background technique [0002] Users' comments on products on e-commerce websites or APPs can help other potential consumers understand the overall picture of the product and help potential consumers make purchase decisions. Many farm apps such as Farm World and Yiyi Farm have adopted the mode of online operation and offline delivery. In order to attract new users and consolidate the trust of old users, it is necessary to provide potential consumers with a receipt display, that is, a product picture of the product received by the user and its evaluation. Therefore, it is very important to study how to find valuable reviews from a large number of product reviews and recommend them to other users. [0003] The recommendation algorithm can filter the products that the user may ...

AI Technical Summary

Problems solved by technology

However, this kind of method ignores the personalization of consumers, does not consider the interests of consumers, and may recommend the same comments to users with different concerns and interests.

Images