计算机技术与发展2024，Vol.34Issue(9)：109-115,7.DOI:10.20165/j.cnki.ISSN1673-629X.2024.0180
融入逻辑规则的知识图谱推荐模型研究
Research on Knowledge Graph Recommendation Models Integrated with Logical Rules
摘要
Abstract
Knowledge graph embedding technology has attracted widespread attention in the field of recommendation systems.Integrating information from structured knowledge graphs into recommendation models can enhance the personalization of recommendations.However,existing knowledge graph recommendation models still face the issue of error propagation due to the inaccuracy of initial data,which leads to incorrect recommendation results.To address this problem,we propose the RR-KGE model,consisting of a knowledge graph embedding module and a recommendation algorithm module.The focus is on the knowledge graph embedding framework,where rule embedding and knowledge graph embedding are jointly learned.Rules provide the model with additional constraints to reduce error propagation.This framework is combined with the recommendation algorithms ALS(Alternating Least Squares)and RNN(Recurrent Neural Network)to obtain more accurate recommendation results.Finally,RR-KGE is compared with different baseline models,and multiple metrics on two datasets demonstrate its superiority over the comparison models,confirming the effectiveness of the recommendation approach.关键词
知识图谱/知识图谱嵌入/逻辑规则/推荐算法/联合学习Key words
knowledge graph/knowledge graph embedding/logical rules/recommendation algorithm/joint learning分类
信息技术与安全科学引用本文复制引用
高文馨,李贯峰,王云丽,胡德洲,李瑞..融入逻辑规则的知识图谱推荐模型研究[J].计算机技术与发展,2024,34(9):109-115,7.基金项目
国家自然科学基金项目(62066038) （62066038）
宁夏自然科学基金项目(2022AAC03026) （2022AAC03026）
宁夏大学研究生创新项目(CXXM202356) （CXXM202356）
