A Bandit Method Using Probabilistic Matrix Factorization in Recommendation

doi:10.1007/s12204-015-1618-7

上海交通大学学报（英文版） ›› 2015, Vol. 20 ›› Issue (5): 535-539.doi: 10.1007/s12204-015-1618-7

A Bandit Method Using Probabilistic Matrix Factorization in Recommendation

TU Shi-tao* (涂世涛), ZHU Lan-juan (朱兰娟)

(Key Laboratory of System Control and Information Processing of Ministry of Education; Department of Automation, Shanghai Jiaotong University, Shanghai 200240, China)

出版日期:2015-10-28 发布日期:2015-10-29
通讯作者: TU Shi-tao (涂世涛) E-mail: tushitao@126.com

A Bandit Method Using Probabilistic Matrix Factorization in Recommendation

TU Shi-tao* (涂世涛), ZHU Lan-juan (朱兰娟)

(Key Laboratory of System Control and Information Processing of Ministry of Education; Department of Automation, Shanghai Jiaotong University, Shanghai 200240, China)

Online:2015-10-28 Published:2015-10-29
Contact: TU Shi-tao (涂世涛) E-mail: tushitao@126.com

摘要/Abstract

摘要： In recommendation system, sparse data and cold-start user have always been a challenging problem. Using a linear upper confidence bound (UCB) bandit approach as the item selection strategy based on the user historical ratings and user-item context, we model the recommendation problem as a multi-arm bandit (MAB) problem in this paper. Enabling the engine to recommend while it learns, we adopt probabilistic matrix factorization (PMF) in this strategy learning phase after observing the payoff. In particular, we propose a new approach to get the upper bound statistics out of latent feature matrix. In the experiment, we use two public datasets (Netfilx and MovieLens) to evaluate our proposed model. The model shows good results especially on cold-start users.

关键词: recommend, matrix factorization, bandit

Abstract: In recommendation system, sparse data and cold-start user have always been a challenging problem. Using a linear upper confidence bound (UCB) bandit approach as the item selection strategy based on the user historical ratings and user-item context, we model the recommendation problem as a multi-arm bandit (MAB) problem in this paper. Enabling the engine to recommend while it learns, we adopt probabilistic matrix factorization (PMF) in this strategy learning phase after observing the payoff. In particular, we propose a new approach to get the upper bound statistics out of latent feature matrix. In the experiment, we use two public datasets (Netfilx and MovieLens) to evaluate our proposed model. The model shows good results especially on cold-start users.

Key words: recommend, matrix factorization, bandit

中图分类号:

TP 181

TU Shi-tao* (涂世涛), ZHU Lan-juan (朱兰娟). A Bandit Method Using Probabilistic Matrix Factorization in Recommendation[J]. 上海交通大学学报（英文版）, 2015, 20(5): 535-539.

TU Shi-tao* (涂世涛), ZHU Lan-juan (朱兰娟). A Bandit Method Using Probabilistic Matrix Factorization in Recommendation[J]. Journal of shanghai Jiaotong University (Science), 2015, 20(5): 535-539.

参考文献 12

[1]	Sarwar B, Karypis G, Konstan J, et al. Itembased collaborative filtering recommendation algorithms[C]//Proceedings of the 10th international conference on World Wide Web. Hong Kong, China:ACM, 2001: 285-295.
[2]	Schein A I, Popescul A, Ungar L H, et al.Methods and metrics for cold-start recommendations[C]//Proceedings of the 25th Annual International ACM SIGIR Conference on Research and Development in Information Retrieval. Tampere, Finland: ACM,2002: 253-260.
[3]	Koren Y, Bell R, Volinsky C. Matrix factorization techniques for recommender systems [J]. Computer,2009, 42(8): 30-37.
[4]	Lee D D, Seung H. Algorithms for non-negative matrix factorization [J]. Advances in Neural Information Processing Systems, 2001, 13: 556-562.
[5]	Jamali M, Ester M. A matrix factorization technique with trust propagation for recommendation in social networks[C]//Proceedings of the 4th ACM Conference on Recommender Systems. Barcelona, Spain:ACM, 2010: 135-142.
[6]	Ma H, Yang H, Lyu M R, et al. Sorec: Social recommendation using probabilistic matrix factorization[C]//Proceedings of the 17th ACM Conference on Information and Knowledge Management. Napa. Valley,California, USA: ACM, 2008: 931-940.
[7]	Macready W G, Wolpert D H. Bandit problems and the exploration/exploitation tradeoff [J]. IEEE Transactions on Evolutionary Computation, 1998,2(1): 2-22.
[8]	Auer P. Using confidence bounds for exploitationexploration trade-offs [J]. The Journal of Machine Learning Research, 2003, 3: 397-422.
[9]	Salakhutdinov R, Mnih A. Probabilistic matrix factorization [C]//Advances in Neural Information Processing Systems. Cambridge, Massachusetts: MIT Press, 2007: 1257-1264.
[10]	Golub G H, Reinsch C. Singular value decomposition and least squares solutions [J]. Numerische Mathematik,1970, 14(5): 403-420.
[11]	Precup D, Sutton R S, Singh S. Eligibility traces for off-policy policy evaluation [C]// Proceedings of 17th International Conference on Machine Learning.San Francisco, CA, USA: Morgan Kaufmann, 2000:759-766.
[12]	Li L, ChuW, Langford J, et al. A contextual-bandit approach to personalized news article recommendation[C]//Proceedings of the 19th International Conference on World Wide Web. Raleish, North Carolina, USA:ACM, 2010: 661-670.

A Bandit Method Using Probabilistic Matrix Factorization in Recommendation

A Bandit Method Using Probabilistic Matrix Factorization in Recommendation

可视化

摘要/Abstract

引用本文

使用本文

参考文献 12

相关文章 4

编辑推荐

Metrics

本文评价

[1]	MAO Qingqing (毛青青), DONG Aihua (董爱华), MIAO Qingying (苗清影), PAN Lu (潘璐). Intelligent Costume Recommendation System Based on Expert System[J]. sa, 2018, 23(2): 227-234.
[2]	LIANG Ye (梁野). Multilingual Financial News Retrieval and Smart Recommendation Based on Big Data[J]. 上海交通大学学报（英文版）, 2016, 21(1): 18-24.
[3]	TANG Song-ze*(唐松泽), XIAO Liang (肖亮), LIU Peng-fei (刘鹏飞). Single Image Super-Resolution Method via Refined Local Learning[J]. 上海交通大学学报（英文版）, 2015, 20(1): 26-31.
[4]	LONG Shun (龙舜), ZHU Wei-heng (朱蔚恒). Mining Evolving Association Rules for E-Business Recommendation[J]. 上海交通大学学报（英文版）, 2012, 17(2): 161-165.