Personalized recommendation method based on multi-view knowledge graph attention network

Abstract

The invention discloses a personalized recommendation method based on a multi-view knowledge graph attention network. The method comprises the following steps: 1) constructing a multi-view knowledge graph; 2) constructing a direct adjacent matrix and an indirect adjacent matrix corresponding to the multi-view knowledge graph according to the multi-view knowledge graph; 3) designing a drawing attention network, and learning multi-view feature information based on an attention mechanism to obtain vector representations of a user and an object; 4) extracting social contact, relation and interaction feature information of the knowledge graph based on the graph attention network; and 5) training and learning the knowledge graph attention network to obtain final vector representation of the user and the object. According to the invention, the accuracy of a recommendation system is improved by learning the multi-view features in the knowledge graph.

Description

technical field [0001] The invention relates to the fields of knowledge graphs, recommendation systems, etc., and in particular provides a personalized recommendation method based on multi-view knowledge graph attention networks. Background technique [0002] With the rapid development of Internet technology, users can choose more and more information, but it also makes users face the problem of "information overload" caused by the large amount of information. How to better provide users with interesting information has become a research hotspot, and the recommendation system is an effective way to solve this problem. Objects are recommended to users to meet their individual needs. [0003] Nowadays, the main problems faced by recommender systems are data sparsity and cold start problem. Data sparsity means that for a large number of users and recommended objects, there is only a small amount of interactive information, making traditional recommendation methods inefficient...

AI Technical Summary

Problems solved by technology

[0007] However, existing recommendation systems based on knowledge graphs often only focus on features from a single perspective, such as object neighborhood information and user social information, and cannot fully capture the feature information in knowledge graphs, resulting in low accuracy of recommendation systems. Low

Images