A Group Recommendation Method Based on Evidence Theory

Abstract

The invention discloses a group recommendation method based on evidence theory, the steps of which include: 1. Using triplets to respectively represent the ratings of users on items and the affiliation relationship between users and groups; 2. Implementing a probability matrix decomposition method, Obtain the predicted score of each member of the group for the project; 3. Define the weight and reliability of each member in the group; 4. For each group, use the evidence reasoning method to fuse the predicted score information of the group members to obtain the group 5. According to the group's predicted scores on the items, sort them from high to low, and select the first W items to finally form the group's recommendation list. The present invention can fully consider the weights of group members when recommending groups, and at the same time give certain reliability to evidence with different weights, and synthesize group recommendation results that satisfy most group members as much as possible by using the method of evidence theory. Thereby effectively improving the effect of group recommendation.

Description

technical field [0001] The invention relates to the technical field of recommendation methods, in particular to a group recommendation method based on evidence theory. Background technique [0002] In recent years, with the development of online communities, a large number of users and information have emerged on major social media, which not only facilitates users, but also brings problems such as information overload. As an effective way to solve the problem of information overload, the recommendation system can help users quickly and effectively find the information they need in massive amounts of data, and meet the individual needs of different users. . [0003] Due to the social nature of human beings, people often like to participate in certain activities together in a certain group form, for example, having dinner with a group of colleagues in the same company after get off work, watching TV programs with family members, and planning travel locations with friends . ...

AI Technical Summary

Problems solved by technology

The minimum regret strategy and the maximum happiness strategy only regard the opinions of the most dissatisfied or most satisfied members in the group as the opinions of the group, which obviously cannot effectively model the group preferences, while the average strategy combines the opinions of the group members when synthesizing group opinions The opinions of all members are treated equally, without taking into account the different contributions of group members in the actual situation. Members are unknown. In fact, different characteristics of different group members will make them play different roles in the process of generating group recommendation results. Existing research methods often ignore this point and reduce the accuracy of group recommendation methods.

Images