《计算机应用研究》|Application Research of Computers

基于二分图网络的总体多样性增强推荐算法

Enhanced algorithm for recommendation aggregate diversity based on bipartite graph networks

免费全文下载 (已被下载 次)  
获取PDF全文
作者 张骏,丁艳辉,金连旭,赵文朋
机构 1.山东师范大学 信息科学与工程学院,济南 250014;2.山东省物流优化与预测工程技术研究中心,济南 250014
统计 摘要被查看 次,已被下载
文章编号 1001-3695(2018)06-1628-03
DOI 10.3969/j.issn.1001-3695.2018.06.006
摘要 针对传统推荐算法过于强调推荐准确率而造成推荐系统长尾现象加剧问题,提出一种基于二分图网络的总体多样性增强推荐算法。首先,利用现有推荐算法生成的预测评分构建用户候选推荐列表,进而构建二分图网络模型;其次,设定项目容量对热门项目的推荐次数予以限制;最后,结合推荐增广路生成最终推荐列表。与现有的推荐多样性增强算法在真实电影评分数据集上进行实验对比,实验结果表明,该算法在保证推荐准确率的同时能有效提高推荐的总体多样性。
关键词 推荐系统;总体多样性;二分图
基金项目 国家自然科学基金青年基金资助项目(61303007)
本文URL http://www.arocmag.com/article/01-2018-06-006.html
英文标题 Enhanced algorithm for recommendation aggregate diversity based on bipartite graph networks
作者英文名 Zhang Jun, Ding Yanhui, Jin Lianxu, Zhao Wenpeng
机构英文名 1.SchoolofInformationScience&Engineering,ShandongNormalUniversity,Jinan250014,China;2.ShandongProvincialLogisticsOptimization&PredictiveEngineeringTechnologyResearchCenter,Jinan250014,China
英文摘要 The traditional recommendation algorithm emphasizes too much on the accuracy of the recommendation, resulting in the “long-tail” phenomenon of the recommendation system intensified.Therefore, this paper proposed an enhanced algorithm for recommendation aggregate diversity based on the bipartite graph networks.Firstly, it constructed candidate user recommendation list based on predictive scores generated by the existing recommendation algorithm, and then constructed the bipartite graph network model.Secondly, set the item-capacity to limit the number of recommendations of popular items.Finally, it gene-rated the final recommendation result in conjunction with the recommendation augmenting path.It compared the experimental with that of the existing recommendation aggregate diversity algorithm on real-world movie rating datasets.The experimental results show that the proposed algorithm can effectively guarantee the accuracy of the recommendation results as well as improve the aggregate diversity of the recommendation.
英文关键词 recommendation system; aggregate diversity; bipartite graph
参考文献 查看稿件参考文献
  [1] Hill W, Stead L, Rosenstein M, et al. Recommending and evaluating choices in a virtual community of use[C] //Proc of ACM SIGCHI Conference on Human Factors in Computing Systems. New York:ACM Press, 1995:194-201.
[2] Resnick P, Varian H R. Recommender systems[J] . Communications of the ACM, 1997, 40(3):56-58.
[3] Kim J E, Kim H G. Music recommendation method with respect to message service:US, US8410347[P] . 2013.
[4] 陈克寒, 韩盼盼, 吴健. 基于用户聚类的异构社交网络推荐算法[J] . 计算机学报, 2013, 36(2):349-359.
[5] 安维, 刘启华, 张李义. 个性化推荐系统的多样性研究进展[J] . 图书情报工作, 2013, 57(20):127-135.
[6] Ziegler C N, Mcnee S M, Konstan J A, et al. Improving recommendation lists through topic diversification[M] //A History of the Comstock Silver Lode & Mines. [S. l. ] :Promontory Press, 1974:22-32.
[7] Niemann K, Wolpers M. A new collaborative filtering approach for increasing the aggregate diversity of recommender systems[C] //Proc of the 19th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York:ACM Press, 2013:955-963.
[8] Adomavicius G, Kwon Y O. Maximizing aggregate recommendation diversity:a graph-theoretic approach[C] //Proc of the 1st International Workshop on Novelty and Diversity in Recommender Systems. 2011:3-10.
[9] Adomavicius G, Kwon Y O. Improving aggregate recommendation diversity using ranking-based techniques[J] . IEEE Trans on Knowledge & Data Engineering, 2012, 24(5):896-911.
[10] 刘慧婷, 岳可诚. 可提高多样性的基于推荐期望的top-N推荐方法[J] . 计算机科学, 2014, 41(7):270-274.
收稿日期 2017/1/15
修回日期 2017/3/14
页码 1628-1630,1667
中图分类号 TP391
文献标志码 A