Online forum recommendation method and system based on dynamic heterogeneous graph and electronic equipment

By adopting a recommendation method based on dynamic heterogeneous graphs in online forums, the dynamic graph neural network model is solved, and the content lag of static recommendation system is realized, personalized and real-time dynamic recommendation is improved, and the user experience is improved.

CN119721202APending Publication Date: 2025-03-28CHINA TELECOM CLOUD TECH CO LTD
0 Cites 0 Cited by

Patent Information

Application Number
CN202411753683.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-12-02
Publication Date
2025-03-28

AI Technical Summary

Technical Problem

The content recommended by the static recommendation system in the prior art is lagging behind and cannot improve the user experience. Especially when user interests and access behaviors in online communities change over time, the static model is difficult to adapt.

Method used

The online forum recommendation method based on dynamic heterogeneous graphs is adopted. By obtaining the attribute information and behavior information of the user's current moment, dynamic heterogeneous graphs are constructed and inputted into the pre-trained dynamic graph neural network recommendation model to update the recommendation results in real time.

Benefits of technology

Real-time capture and adaptation to user interests and community dynamics is achieved, providing users with personalized and real-time dynamic recommendations, and improving the accuracy and user experience of recommendations.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The embodiment of the invention provides an online forum recommendation method and system based on a dynamic heterogeneous graph and electronic equipment. The method comprises the steps that attribute information of a user at the current moment, behavior information of accessing post nodes and behavior information of sub forum nodes to which the post nodes belong are obtained; according to the attribute information of the user at the current moment, the behavior information of accessing the post node and the behavior information of the sub-forum node to which the post node belongs, constructing a behavior composition graph; inputting the behavior composition graph into a pre-trained recommendation model, and obtaining to-be-recommended sub forums and posts output by the recommendation model; the pre-trained recommendation model is a model trained in a time sequence by adopting an asynchronous training mode according to historical data of a user. The method solves the problem that the user experience cannot be improved due to the fact that the content recommended by a static recommendation system is lagged in the prior art.
Need to check novelty before this filing date? Find Prior Art