Collaborative filtering method based on item class scoring of users

Abstract

The invention discloses a collaborative filtering method based on item class scoring of users. According to the method, scoring of the item classes is introduced on the basis of a characteristic matrix, and the preference of the users on the item classes on the basis of the access frequency of the users is obtained according to subjective scores of the users, that is, real preference of the usersis obtained; the final similarity of the users is obtained according to the real preference of the users and the score similarity. By conducting test verification on a real data set and introducing the scores on the user characteristic matrix, the user preference can be reflected more accurately, the real scores of the users are predicted, and the accuracy of a recommendation system is improved. The method overcomes the defect that existing methods are used for recognizing interest and preference of the users only according to user objective behaviors and cannot truly reflect the user preference from the subjective aspect, and then the quality of the recommendation system is effectively improved.

Description

technical field [0001] The invention relates to a recommendation method, in particular to a collaborative filtering method based on users' ratings on items. The algorithm is used to make recommendations for users based on the user's real interest preference. Background technique [0002] Currently, the collaborative filtering recommendation method is one of the most widely used and successful recommendation techniques in the recommendation system. It is based on user interests, finds similar users of the specified user in the user group, and integrates the evaluation of a certain information by these similar users to form a system prediction of the specified user's preference for this information. It establishes the user's personal interest model based on the user's personal information and user's historical feedback, and then recommends information of interest to each user. [0003] Collaborative filtering methods can be divided into user-based and item-based collaborative...

AI Technical Summary

Problems solved by technology

The shortcomings of the collaborative filtering method that judges user preferences from the number of times users visit items are as follows: 1) It is not comprehensive to analyze user interests only based on the number of visits

2) Users do not make full use of the ratings of items

Images