Multi-user recommendation system based on knowledge graph path reasoning

Abstract

The invention discloses a multi-user recommendation system based on knowledge graph path reasoning. The multi-user recommendation system comprises a knowledge graph construction module, a path reasoning module and a score prediction module, the knowledge graph construction module obtains user interaction historical data, constructs a knowledge graph G and transmits the knowledge graph G to the path reasoning module; the path reasoning module generates a relation path between a plurality of users and a target project according to the knowledge graph G, and transmits the relation path to the score prediction module; and the score prediction module evaluates and predicts the plurality of received relation paths, and outputs a project recommendation list to the user according to an evaluation and prediction result. According to the method, more abundant user interest latent semantic information is mined by constructing the knowledge graph of the users and the items, and potential common preferences of the multiple users are predicted through combined reasoning of multiple paths on the knowledge graph; pooling aggregation operation with an attention mechanism is utilized between different paths, different user preference degrees are distinguished, and meanwhile a recommendation list maximizing interest preference and diversification is given.

Description

technical field [0001] The invention relates to the field of user recommendation, in particular to a multi-user recommendation system based on knowledge graph path reasoning. Background technique [0002] Most of the recommendation technologies studied in academia and industry are personalized recommendation systems for a single user, but the recommended objects are often groups of more than one person, and then group recommendation technology was born. Groups with similar interests form groups, and then the preferences of members in the group are integrated to achieve group aggregation. For example, users who frequently browse or purchase similar products are aggregated into a group by mining user interaction history. However, in real life, there are naturally existing groups such as families, classes, companies, etc. The interest correlation between these groups may be very small. At the same time, in some cases, these groups can only be selected through the same device ID...

AI Technical Summary

Problems solved by technology

[0007] 3) There is a lack of social relationships between groups, how to use the interaction history of existing users and items to mine more effective information;

Images