基于置信度传播的协同过滤推荐算法  

Collaborative Filtering Recommendation Algorithm Based on Belief Propagation

作　　者：潘燕梅 PAN Yan-mei(Industral and Commercial Department,Huangshan Vocational And Technical College,Huangshan 245000,China)

机构地区：[1]黄山职业技术学院工业与财贸系,安徽黄山245000

出　　处：《通化师范学院学报》2021年第2期109-114,共6页Journal of Tonghua Normal University

摘　　要：随着电子商务的发展,个性化推荐系统已成为研究热点.该文在传统协同过滤推荐算法的基础上引入了重要性程度因子,并提出了基于置信度传播的协同过滤推荐算法.算法通过引入用户重要性程度因子和项目重要性程度因子,基于通信中低密度奇偶校验码(LDPC码)的置信度传播译码算法迭代更新用户和项目的重要性程度因子,从而最终计算得到目标用户对项目的预测评分.Matlab仿真结果表明,基于置信度传播的协同过滤推荐算法相较于传统算法精确度可提升5.09%.With the development of e-commerce,personalized recommendation system has received a lot of attention.In this paper,the importance degree factor is introduced on the basis of the traditional collab⁃orative filtering recommendation algorithm,and a collaborative filtering recommendation algorithm based on belief propagation is proposed.The importance factors of user factors and items are introduced firstly in the user similarity score calculation and prediction process,then use the belief propagation(BP)decod⁃ing algorithm of low density parity check code(LDPC code)updating importance factor.The experimen⁃tal results show that the accuracy of the proposed collaborative filtering recommendation algorithm based on belief propagation outperforms 5.09%compared with the traditional collaborative filtering algorithm.

关 键 词：电子商务 个性化推荐 协同过滤 置信度传播 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]