Collaborative filtering recommendation method based on user preferences

Abstract

The invention discloses a collaborative filtering recommendation method based on user preferences. The method includes the following steps that according to a threshold value p, users are divided into with high-score preference users and low-score preference users; an NHSM similarity measuring method is adopted to calculate the similarity between any two users, and a similarity matrix among the users is obtained; target users Ua are set, a same-preference user group of the target user Ua is s, a different-preference user group of the target user Ua is d, the similarity between the target users Ua and other users Ub is calculated; a new similarity matrix Newsim among the users is calculated; scores which the target users Ua give for un-scored items by the target users Ua are predicted. Compared with the prior art, the collaborative filtering recommendation method based on the user preferences has the following advantages that a collaborative filtering recommendation model is constructed on the basis of the user preferences and the contingency of predicted results is reduced. Test results show that compared with other recommendation methods, through the collaborative filtering recommendation method based on the user preferences, the similarity among users can be more accurately measured, and the recommendation quality is improved.

Description

technical field [0001] The invention relates to the field of Internet personalized recommendation, in particular to a collaborative filtering recommendation method based on user preference. Background technique [0002] The rapid development of Internet technology has increased the capacity of web information, how to provide users with valuable information has become a problem that e-commerce must face. The recommendation system understands and grasps the user's preferences by learning the user's behavior, so that it can recommend more targeted content to the user that they may be interested in. Collaborative filtering is a popular technology in recommender systems, which can effectively solve the problem of information overload, help users quickly find valuable information, and filter useless information. The basic idea is: according to the user's purchase records, rating records, browsing records and annotations, etc., recommend the information needed by the user or predi...

AI Technical Summary

Problems solved by technology

However, users have preference behaviors when scoring, and there are situations where users score low or high, and users with similar interests also have different scores, which will lead to deviations in the calculation of similarity between users, which affects the quality of recommendations

Images