融合多信息的个性化推荐模型  

Personalized Recommendation Model with Multiple Information Fusion

作　　者：沈杰 乔少杰[1,2] 韩楠 元昌安 许源平[1] 覃晓 王珏岚[5] SHEN Jie;QIAO Shaojie;HAN Nan;YUAN Changan;XU Yuangping;QIN Xiao;WANG Juelan(Chengdu University of Information Technology,Chengdu 610225,China;Automatic Software Generation and In telligent Service Key Laboratory of Sichuan Province,Chengdu 610225,China;Guangxi College of Education,Nanning 530007,China;Nanning Normal University,Nanning 530299,China;Sichuan Academy of Medical Sciences&Sichuan Provincial People's Hospital,School of Medicine,University of Electronic Science and Technology of China,Chengdu 610072,China)

机构地区：[1]成都信息工程大学,成都610225 [2]软件自动生成与智能服务四川省重点实验室,成都610225 [3]广西教育学院,南宁530007 [4]南宁师范大学,南宁530299 [5]四川省医学科学院·四川省人民医院,电子科技大学附属医院,成都610072

出　　处：《重庆理工大学学报（自然科学）》2021年第3期128-138,共11页Journal of Chongqing University of Technology：Natural Science

基　　金：四川省教育厅人文社会科学重点研究基地四川景观与游憩研究中心科研项目(JGYQ2018010);成都市新冠肺炎防控科技项目(2020-YF05-00058-SN);广西自然科学基金项目(2018GXNSFDA138005);国家自然科学基金项目(61802035,61772091,61962006,U2001212,62072311);四川省科技计划项目(2021JDJQ0021,2020YFG0153,20YYJC2785,2020YJ0481,2020YFS0466,2020YJ0430,2020JDR0164,2019YFS0067,2020YFS0399);四川高校科研创新团队建设计划项目(18TD0027)。

摘　　要：为了解决协同过滤推荐系统的数据稀疏与冷启动问题,通常利用辅助信息来提高推荐系统的整体性能。当前大多数社交媒体网站和电子商务系统都允许用户发表文本评论,以及对项目(如商户、电影、商品等)进行评分。为了更加有效地融合多种数据信息,解决数据稀疏的问题,提高推荐算法的准确性,构建了一个基于用户-项目历史交互数据源融合知识图谱的模型,提出了基于用户-项目历史评论的深度学习算法,将2种算法动态融合,利用随机梯度下降方法进行模型求解,为用户提供更精准的个性化推荐服务。实验结果表明:相比于已有典型推荐算法,所提模型取得更好的推荐效果,并且可以有效地解决数据稀疏带来的推荐准确性降低的问题。To cope with the problem of data sparseness and cold start in collaborative filtering recommendation systems,auxiliary information is usually used to improve the overall performance of recommendation systems.Currently,most social media websites and e-commerce systems allow users to post textual reviews to express their personal opinions towards the purchased items(e.g.,customers,movies,goods),along with a rating score indicating their preferences.In order to effectively integrate multiple data information,solve the problem of data sparseness,improve the accuracy of recommendation algorithms,a model based on user-item historical interaction data and knowledge graph is constructed,and a deep learning algorithm based on user-item historical reviews is proposed.These two algorithms are dynamically integrated,and the stochastic gradient descent algorithm is applied to solve the problem in order to make more accurate personalized recommendation for users.Extensive experimental results on real datasets showed that the proposed model achieves better recommendation results when compared with existing typical recommendation algorithms and can effectively handle the problem of reduced recommendation accuracy caused by data sparseness.

关 键 词：推荐系统 知识图谱 深度学习 动态融合 随机梯度下降 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]