Recommendation method and system based on knowledge-aware hypergraph neural network

Abstract

The invention discloses a recommendation method based on a knowledge-aware hypergraph neural network, which comprises the following steps: step 1, constructing a user hypergraph, and initializing a hyperedge by taking an article having an interaction relationship with a user as an entity node; constructing an article hypergraph, wherein the initial hyperedge takes an article having an interaction relationship with any user having the interaction relationship with the article as an entity node; step 2, convolution calculation is performed on the user hypergraph and each article hypergraph; and step 3, calculating the inner product of the user and all the articles to obtain the interaction score of the user and all the articles. The method has the beneficial effect that the auxiliary information of the articles is integrated into the vector representation of the user, so that the articles of which the user is more likely to have the interaction relationship can be selected from the numerous articles. The invention discloses a recommendation system based on a knowledge-aware hypergraph neural network. The recommendation system comprises: a data capture module; a knowledge perception hypergraph construction module; a domain convolution module; a hyperedge convolution module; and a prediction module. The method has the beneficial effects of simple operation and accurate prediction.

Description

technical field [0001] The invention relates to the technical field of knowledge perception application. More specifically, the present invention relates to a recommendation method and system based on a knowledge-aware hypergraph neural network. Background technique [0002] With the rapid development of the Internet, recommender systems are widely deployed to alleviate the effects of information overload. A traditional recommendation technique is collaborative filtering, which assigns representation vectors based on user and item identities, and then models their interactions through specific operations such as inner products or neural networks. However, in the absence of auxiliary information, these collaborative filtering-based methods usually suffer from sparsity and cold-start problems. To address these issues, various types of auxiliary information have been explored to improve recommendation performance, such as item attributes, item reviews, and users' social netwo...

AI Technical Summary

Problems solved by technology

However, these methods still suffer from two limitations: (1) Higher-order correlations among users, items, and entities in knowledge graphs are essential for data modeling

These methods mainly apply graph neural networks to enrich the representation of target nodes by recursively aggregating their original neighbors in the knowledge graph, and thus have limitations in modeling high-order correlations between target nodes and non-original neighbors.

Furthermore, the graph structure employed by existing methods has constraints on modeling and handling higher-order dependencies, since graphs can only represent pairwise connections

Images