当前位置>主页 > 期刊在线 > 信息技术 >

信息技术23年1期

基于社区结构的推荐系统中冷启动问题研究
曹珂崯,张天舒,孙娟,彭二帅
(江苏电子信息职业学院 计算机与通信学院,江苏 淮安 223001)

摘  要:由于推荐系统中数据稀疏性和冷启动问题日益严重,传统的算法无法有效地解决这些问题,现有的改进算法由于稳定性差,仍然需要预先确定具体的参数。文章提出了一种基于社区结构的冷启动算法以改善推荐系统中的冷启动问题,通过计算用户和项目节点的相似度投影二分网络,利用改进的 Louvain 算法对投影单模式网络进行社区检测,使新记录更新到社区,然后进行对用户社区组推荐。与其他冷启动算法相比,该算法在推荐准确率和运行效率有明显提升。


关键词:社区检测;冷启动;二分投影网络;推荐系统



DOI:10.19850/j.cnki.2096-4706.2023.01.007


中图分类号:TP301.6;TP391                            文献标识码:A                                文章编号:2096-4706(2023)01-0030-04


Research on Cold Start Problem in Recommendation System Based on Community Structure

CAO Keyin, ZHANG Tianshu, SUN Juan, PENG Ershuai

(Department of Computer and Communication, Jiangsu Vocational College of Electronics and Information, Huaian 223001, China)

Abstract: Due to the increasingly serious problems of data sparsity and cold start in recommendation systems, traditional algorithms cannot effectively solve these problems, and the existing improved algorithms still need to pre-determine specific parameters due to their poor stability. In this paper, a cold-start algorithm based on community structure is proposed to improve the cold-start problem in recommendation systems. By calculating the similarity bipartite projection network between users and project nodes, the improved Louvain algorithm is used to detect the community for the projected single-mode network. It updates the new record to the community and then makes recommendations to the user community groups. Compared with other cold-start algorithms, the algorithm has a significant improvement in recommendation accuracy and operating efficiency.

Keywords: community detection; cold start; bipartite projection network; recommendation system


参考文献:

[1] 于鹏华 . 数据数量与质量敏感的推荐系统若干问题研究[D]. 杭州:浙江大学,2016.

[2] FUNK S. Netflix update:try this at home [EB/OL].(2006-12-11).http://sifter.org/simon/journal/20061211.html.

[3] 朱传亮 . 基于社交关系和社区结构的协同过滤推荐算法研究 [D]. 重庆:重庆邮电大学,2019.

[4] 张凯涵,梁吉业,赵兴旺,等 . 一种基于社区专家信息的协同过滤推荐算法 [J]. 计算机研究与发展,2018,55(5):968-976.

[5] 张川 . 基于内容和用户行为的个性化微博推荐算法研究与实现 [D]. 北京:北京邮电大学,2018.

[6] SHI L ,ZHAO W X ,SHEN Y D. Local RepresentativeBased Matrix Factorization for Cold-Start Recommendation [J].Acm Transactions on Information Systems,2017,36(2):1-28.

[7] 周超,孙英华,熊化峰,等 . 基于用户和项目双向聚类的协同过滤推荐算法 [J]. 青岛大学学报:自然科学版,2018,31(1):55-60.

[8] BLONDEL V D,GUILLAUME J,LAMBIOTTE R, et al.Fast unfolding of communities in large networks [J].Journal of Statistical Mechanics:Theory and Experiment,2008(10):1-12.

[9] 陈东明,严燕斌,黄新宇,等 . 基于二分网络社团划分的推荐算法 [J]. 东北大学学报:自然科学版,2018,39(8):1103-1107.

[10] 夏玮,杨鹤标 . 改进的 Louvain 算法及其在推荐领域的研究 [J]. 信息技术,2017(11):125-128.

[11] BI X,QU A,WANG J,et al. A Group-Specific Recommender System [J].Journal of the American Statal Association,2017,112(519):1344-1353.

[12] 王金 . 基于 SVD++ 的协同过滤群组推荐算法研究 [D].金华:浙江师范大学,2018.


作者简介:曹珂崯(1994—),男,汉族,山东鄄城人,助教,硕士,研究方向:网络安全、人工智能。