基于用户偏好加权的混合网络推荐算法

doi:10.6040/j.issn.1671-9352.3.2014.232

山东大学学报（理学版） ›› 2015, Vol. 50 ›› Issue (09): 29-35.doi: 10.6040/j.issn.1671-9352.3.2014.232

基于用户偏好加权的混合网络推荐算法

张新猛^1,2, 蒋盛益^1,2, 张倩生^2,3, 谢柏林^1,2, 李霞²

1. 广东外语外贸大学信息学院, 广东广州 510006;
2. 语言工程与计算广东省社会科学重点实验室, 广东广州 510006;
3. 广东外语外贸大学金融学院, 广东广州 510006

收稿日期:2015-03-03 修回日期:2015-07-22 出版日期:2015-09-20 发布日期:2015-09-26
作者简介:张新猛(1974-),男,硕士,副教授,主要研究方向为社会网络和推荐系统.E-mail:javad0902@163.com
基金资助:
国家自然科学基金资助项目(61202271);国家社会科学基金资助项目(13CGL130);教育部人文社会科学研究青年基金资助项目(13YJCZH258);广东省自然科学基金资助项目(S2012040007184,S2013010013050);广东省普通高校科技创新项目(2013KJCX0069, 2012KJCX0049)

Hybrid recommendation by combining network-based algorithm and user preference

ZHANG Xin-meng^1,2, JIANG Sheng-yi^1,2, ZHANG Qian-sheng^2,3, XIE Bo-lin^1,2, LI Xia²

1. Cisco School of Informatics, Guangdong University of Foreign Studies, Guangzhou 510006, Guangdong, China;
2. Social Science Key Laboratory of Language Engineering and Computing of Guangdong Province, Guangzhou 510006, Guangdong, China;
3. School of Finance, Guangdong University of Foreign Studies, Guangzhou 510006, Guangdong, China

Received:2015-03-03 Revised:2015-07-22 Online:2015-09-20 Published:2015-09-26

摘要/Abstract

摘要： 基于热传导或物质扩散理论的推荐算法首先利用网络结构得到对象间推荐关系,然后根据对象间关系预测用户喜欢的对象,而忽略了用户偏好。为了弥补这个缺陷,根据用户已选择对象的标签,利用TF-IDF方法构建用户偏好模型,以用户在预测对象标签上的平均偏好作为对该对象的偏好程度,采用加权方法与现有基于网络推荐算法混合运算。经在基准数据集MovieLens上测试表明,通过与目前效果最好的几种基于网络推荐算法进行加权混合运算,推荐结果在推荐精度、个性化、多样化等多种评价指标方面均比原有算法有明显提高。

关键词: TF-IDF, 个性化推荐, 标签, 基于网络推荐

Abstract: Recommendation algorithms based on heat conduction or mass diffusion first obtain the relationship between objects according to network structure, then predict the user's favorite objects based on these relationships, but these algorithms ignore user's preference. In order to overcome this defect, the TF-IDF approach was used to construct user's preference according to the tags contained in the objects selected by user, and the mean of preference of object's tags was taken as the preference of the object, then a hybrid recommendation model was proposed by combining network-based algorithm and the user preference model. The benchmark datasets, MovieLens, was used to evaluate our algorithm, and the experimental results demonstrate that hybrid algorithm can significantly improve accuracy, diversification and personalization of recommendations.

Key words: TF-IDF, network-based recommendation, personalized recommendation, tag

中图分类号:

TP393

张新猛, 蒋盛益, 张倩生, 谢柏林, 李霞. 基于用户偏好加权的混合网络推荐算法[J]. 山东大学学报（理学版）, 2015, 50(09): 29-35.

ZHANG Xin-meng, JIANG Sheng-yi, ZHANG Qian-sheng, XIE Bo-lin, LI Xia. Hybrid recommendation by combining network-based algorithm and user preference[J]. JOURNAL OF SHANDONG UNIVERSITY(NATURAL SCIENCE), 2015, 50(09): 29-35.

参考文献

[1] ZHOU T, REN J, MEDO M, et al. Bipartite network projection and personal recommendation[J]. Physical Review E, 2007, 76(4):6116-6123.
[2] ZHOU T, JIANG L L, SU R Q. Effect of initial configurationon network-based recommendation[J]. Europhys Lett, 2008, 81(5):8004-8008.
[3] PAN X, DENG G S, LIU J G. Weighted bipartite network and personalized recommendation[J]. Physics Procedia, 2010, 3(5):1867-1876.
[4] LIU J G, ZHOU T, WANG B H, et al. Effects of user tastes on personalized recommendation[J]. International Journal of Modern Physics C, 2009, 20(12):1925-1932.
[5] ZHOU T, SU R Q, LIU R R, et al. Accurate and diverse recommendations via eliminating redundant correlations[J]. New J Phys, 2009, 11(12):3008-3026.
[6] ZHANG Y C, BLATTNER M, YU Y K. Heat conduction process on community networks as a recommendation model[J]. Phys Rev Lett, 2007, 99(15):4301-4305.
[7] LIU J G, ZHOU T, GUO Q. Information filtering via biased heat conduction[J]. Phys Rev E, 2011, 84(3):7101-7105.
[8] QIU T, WANG T T, ZHANG Z K, et al. Heterogeneity involved network-based algorithm leads to accurate and personalized recommendations[J]. Physics and Society, 2013, arXiv:1305.7438vl.
[9] ZHOU T, KUSCSIK Z, LIU J G, et al. Solving the apparent diversity accuracy dilemma of recommender systems[C]//Proceedings of the National Academy of Sciences of the United States of America. Washington: Natl Acad Sciences, 2010, 107:4511-4515.
[10] SCOTT A G, BERNARDO A H. Usage patterns of collaborative tagging systems[J]. Journal of Information Science, 2006, 32(2):198-208.
[11] ZHANG Z K, ZHOU T, ZHANG Y C. Tag-aware recommender systems: a state-of-the-art survey[J]. Journal of Computer Science and Technology, 2011, 26(5):767-777.
[12] ZHANG Z K, LIU C, ZHANG Y C, et al. Solving the cold-start problem in recommender systems with social tags [J]. Europhysics Letters, 2010, 92(2):8002-8010.
[13] MICHAEL J P, DANIEL B. Content-based recommendation systems[J]. Lecture Notes in Computer Science, 2007, 4321:325-341.
[14] CANTADOR I, BELLOGN A, VALLET D. Content-based recommendation in social tagging systems[C]//Proceedings of RecSys'10.New York: ACM, 2010:237-240.
[15] JIANG Shengyi, SONG Xiaoyu, WANG Hui, et al. A clustering-based method for unsupervised intrusion detections[J]. Pattern Recognition Letters, 2006, 27(7):802-810.
[16] BURKE R. Hybrid Recommender systems: survey and experiments [J]. User Model User-Adap Interact, 2007, 12(4):331-370.
[17] JOACHIMS T. A probabilistic analysis of the rocchio algorithm with TFIDF for text categorization[C]//Proceedings of the 14th International Conference on Machine Learning. New York: ACM, 1997:143-151.

多维度评价

Viewed

Full text

Abstract

Cited

Shared

Discussed

基于用户偏好加权的混合网络推荐算法

Hybrid recommendation by combining network-based algorithm and user preference

PDF (PC)

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 7

多维度评价

本文评价

推荐阅读 0

[1]	杜漫,徐学可,杜慧,伍大勇,刘悦,程学旗. 面向情绪分类的情绪词向量学习[J]. 山东大学学报（理学版）, 2017, 52(7): 52-58.
[2]	张聪,裴家欢,黄锴宇,黄德根,殷章志. 基于语义图优化算法的中文微博观点摘要研究[J]. 山东大学学报（理学版）, 2017, 52(7): 59-65.
[3]	杨震,司书勇,李超阳. 基于用户隐式兴趣模型的信息推荐[J]. 山东大学学报（理学版）, 2017, 52(1): 15-22.
[4]	马宇峰, 阮彤. 基于LDA及标签传播的实体集合扩展[J]. 山东大学学报（理学版）, 2015, 50(03): 20-27.
[5]	王少鹏, 彭岩, 王洁. 基于LDA的文本聚类在网络舆情分析中的应用研究[J]. 山东大学学报（理学版）, 2014, 49(09): 129-134.
[6]	刘璇1,许洁萍1*,陈捷2. 以Web标签为基础的相似歌曲研究[J]. J4, 2012, 47(5): 53-58.
[7]	刘健1,尹春霞2*,原福永3. 基于非结构化P2P网络用户模型的协同过滤推荐机制[J]. J4, 2011, 46(5): 28-33.