期刊文献+

基于用户信任及推荐反馈机制的社会网络推荐模型 预览 被引量:2

SOCIAL NETWORK RECOMMENDATION MODEL BASED ON USER TRUST AND RECOMMENDATION FEEDBACK MECHANISM
在线阅读 下载PDF
收藏 分享 导出
摘要 社会网络包括以兴趣为核心的兴趣网络和以信任为核心的信任网络.如何利用社会网络中用户信任与兴趣相似的好友的项目数据来扩展用户本身的项目数据集,缓解用户数据稀疏性,利用目标用户的好友的项目评分数据为其产生推荐,是研究的重点.和传统的推荐方法相比,提出一种改进模型S I M T M ( Similar a n d Trust M o d e l ) 来提供用户更加高效的推荐体验.该模型融合用户兴趣度和信任度作为初始亲密程度,根据融合后的好友网络进行推荐,同时根据推荐反馈,来不断地优化用户的项目评分数据集,使得亲密的用户好友更加亲密,过滤掉用户的普通好友,优化用户之间的兴趣和信任关联;并重新计算用户之间的亲密程度形成融合用户与其好友的融合网络,直至前后两次根据亲密程度得到的推荐结果相近,根据得到的最优的亲密程度构建融合网络来进行推荐.实验结果表明,该模型在数据稀疏的情况下,能有效提高用户推荐的准确率和覆盖率. Social networks include the interest network taking the interest as core and the trust network taking the trust as core. The research focus of this paper is that how to use the projects data of the friends in social networks with similar trust and interest to expand the project dataset of user's own,to alleviate the sparsity of user data,and to use the data of project rating score of target user's friends to generate recommendation for it. Compared with traditional recommendation methods,the paper presents an improved SIMTM( Similar and Trust Model),which can provide more efficient recommendation experience. The model fuses interest and confidence as the initial intimacy,and makes recommendation according to the fused networks of friends,at the same time it constantly optimises the project rating score dataset according to the recommended feedbacks,this makes user's close friends be more intimate while filtering out user's ordinary friends,and optimises the association of interest and trust between user,moreover it re-calculates the intimacy degree between users to form a fusion network which fuses the user and user's friends until the twice recommendation results before and the after derived from intimacy degree are close,and then constructs the fusion network based on the derived optimal intimacy degree for recommendation. Experimental results show that,the model can effectively improve the accuracy and coverage of recommendation of users,especially in the case of data sparsity.
作者 翟鹤 刘柏嵩 Zhai He;Liu Baisong;College of Information Science and Engineering,Ningbo University;
出处 《计算机应用与软件》 CSCD 2016年第11期258-262,共5页 Computer Applications and Software
关键词 社会网络 兴趣网络 信任网络 融合网络 推荐反馈 信任更新 Social network Interested network Trust network Fusion network Recommendation feedback Trust update
作者简介 翟鹤,硕士生,主研领域:数据挖掘. 刘柏嵩,研究员.
  • 相关文献

参考文献4

二级参考文献94

  • 1O'Donovan J, Smyth B. Trust in Recommender Systems[C]//Proc. of IUI'05. San Diego, California, USA: [s. n.],2005: 167-174. 被引量:1
  • 2Massa P, Bhattacharjee B. Using Trust in Recommender Systems: An Experimental Analysis[C]//Proc. of the 2nd International Conference on Trust Management. Oxford, UK: [s. n.], 2004: 104- 110. 被引量:1
  • 3Goldberg D, Nichols D, Oki B M, et al. Using Collaborative Filtering to Weave an Information Tapestry[J]. Communications of the ACM, 1992, 35(12): 61-70. 被引量:1
  • 4Herlocker J L, Konstan J A, Terveen L G, et al. Evaluatin Collaborative Filtering Recommender Systems[J]. ACM Transactions on Information Systems, 2004, 22(1): 5-53. 被引量:1
  • 5Guo S, Wang M, Leskovoc J. The role of social networks in on- line shopping : information passing, price of trust, and consumer choice[ C]. Proceedings of the 12th ACM Conference on Electron- ic Commerce, ACM Press, New York, 2011 : 157-166. 被引量:1
  • 6Guy I, Carmel D. Social recommender systems [ C ]. Proceedings of the 20th International Conference Companion on World Wide Web, 2011. 被引量:1
  • 7Sarwar B M. Sparsity, scalability, and distribution in recommender systems[ D]. Minneapolis, USA: University of Minnesota, 2001. 被引量:1
  • 8Zhou Tao. The ten challenges of personalized recommendation [ J ]. Communications of the China Computer Federation, 2012, 8 (7) : 48-61. 被引量:1
  • 9Movielens data sets[EB/OL], http://www. grouplens. org/node, 2011-08 -24. 被引量:1
  • 10Datasets [ EB/OL ]. http ://it. ii. uam. es/hetrec2011/datasets.ht- ml,2011-10-23. 被引量:1

共引文献36

同被引文献13

引证文献2

二级引证文献1

投稿分析

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部 意见反馈