计算机应用研究2024，Vol.41Issue(1)：65-71,7.DOI:10.19734/j.issn.1001-3695.2023.05.0280
融合交互强度的优化社交推荐算法
Improving social recommendation algorithm via incorporating interaction strength
摘要
Abstract
Existing social recommendation algorithms ignore the investigation on the association between rating information and social information.To address this issue,this paper proposed a social recommendation algorithm which incorporated interaction strength.Firstly,it utilized social information and rating data to enrich the social matrix by combining two kinds of similari-ties.Secondly,it defined the interaction strength to represent complex relationship between users.Finally,it introduced a new objective function to learn features of users and items for personalized recommendation using two types of associations,namely the association between interaction strength and social relationships,and the association between features of users and partici-pation features of group which users belonged to.Experimental results on three real-world datasets indicate that the proposed algorithm shows significant improvement in terms of recommendation prediction accuracy compared with existing baseline mo-dels.Furthermore,the proposed algorithm behaves good robustness in learning latent features for users with different number of ratings.Based on the above observations,it can infer that incorporating interaction strength is beneficial to enhancing social recommendation performance and improving users'experience.关键词
社交化推荐/交互强度/群体参与矩阵/矩阵分解Key words
social recommendation/interaction strength/group participation matrix/matrix factorization分类
信息技术与安全科学引用本文复制引用
周璐鑫,李曼,蒋明阳,张雷..融合交互强度的优化社交推荐算法[J].计算机应用研究,2024,41(1):65-71,7.基金项目
国家自然科学基金资助项目(62276034) （62276034）
重庆市教育委员会科学技术研究项目(KJQN202100712) （KJQN202100712）
重庆市研究生科研创新项目(CYS22429) （CYS22429）
