JZUS - Journal of Zhejiang University SCIENCE

ENGINEERING Information Technology & Electronic Engineering

Accepted manuscript available online (unedited version)

Preference transfer model in collaborative filtering for implicit data

Author(s): Bin Ju, Yun-tao Qian, Min-chao Ye
Affiliation(s): Institute of Artificial Intelligence, College of Computer Science and Technology, Zhejiang University, Hangzhou 310027, China; more
Corresponding email(s): jubin_hz@163.com, ytqian@zju.edu.cn, yeminchao@126.com
Key Words: Recommender systems, Collaborative filtering, Preference transfer model, Cross domain, Implicit data

Share this article to： More \|Next Paper >>>

Bin Ju, Yun-tao Qian, Min-chao Ye. Preference transfer model in collaborative filtering for implicit data[J]. Frontiers of Information Technology & Electronic Engineering,in press.https://doi.org/10.1631/FITEE.1500313

@article{title="Preference transfer model in collaborative filtering for implicit data",
author="Bin Ju, Yun-tao Qian, Min-chao Ye",
journal="Frontiers of Information Technology & Electronic Engineering",
year="in press",
publisher="Zhejiang University Press & Springer",
doi="https://doi.org/10.1631/FITEE.1500313"
}

%0 Journal Article
%T Preference transfer model in collaborative filtering for implicit data
%A Bin Ju
%A Yun-tao Qian
%A Min-chao Ye
%J Frontiers of Information Technology & Electronic Engineering
%P 489-500
%@ 2095-9184
%D in press
%I Zhejiang University Press & Springer
doi="https://doi.org/10.1631/FITEE.1500313"

TY - JOUR
T1 - Preference transfer model in collaborative filtering for implicit data
A1 - Bin Ju
A1 - Yun-tao Qian
A1 - Min-chao Ye
J0 - Frontiers of Information Technology & Electronic Engineering
SP - 489
EP - 500
%@ 2095-9184
Y1 - in press
PB - Zhejiang University Press & Springer
ER -
doi="https://doi.org/10.1631/FITEE.1500313"

Abstract
Chinese Summary
Academic Network
Reviewer Comment

Abstract: Generally, predicting whether an item will be liked or disliked by active users, and how much an item will be liked, is a main task of collaborative filtering systems or recommender systems. Recently, predicting most likely bought items for a target user, which is a subproblem of the rank problem of collaborative filtering, became an important task in collaborative filtering. Traditionally, the prediction uses the user item co-occurrence data based on users’ buying behaviors. However, it is challenging to achieve good prediction performance using traditional methods based on single domain information due to the extreme sparsity of the buying matrix. In this paper, we propose a novel method called the preference transfer model for effective cross-domain collaborative filtering. Based on the preference transfer model, a common basis item-factor matrix and different user-factor matrices are factorized. Each user-factor matrix can be viewed as user preference in terms of browsing behavior or buying behavior. Then, two factor-user matrices can be used to construct a so-called ‘preference dictionary’ that can discover in advance the consistent preference of users, from their browsing behaviors to their buying behaviors. Experimental results demonstrate that the proposed preference transfer model outperforms the other methods on the Alibaba Tmall data set provided by the Alibaba Group.

Abstract: The authors propose to transfer data from browsing history of users into user-item matrix including bought items in order to improve the prediction accuracy of collaborative filtering schemes. Such filtering systems predict whether a user will buy an item or not. The authors also perform real data-based experiments to evaluate their proposed scheme. The paper is very clear and very well written. The paper focuses on an interesting problem.

基于兴趣转移模型的协同过滤算法

目的：由于单一域中的用户-物品关系数据的极度稀疏性，很容易遇到预测的“天花板”。为提高隐性关系数据的预测效果，把来自浏览域的用户-物品频次矩阵与来自购物域的用户-物品频次矩阵进行整合，能够提早预测到用户兴趣的时间发生点，从而提高推荐系统的预测精度。
创新点：采用多任务非负矩阵分解框架。不同于前人基于高斯分布先验的矩阵分解模型，本文基于泊松分布先验，提出一种基于用户兴趣因子转移的跨域的非负矩阵分解模型。
方法：首先，“先看后买”是人类购物的一般行为。来自浏览行为背后的用户兴趣一定早于购物行为背后的用户兴趣发生，因此通过概率图模型对矩阵进行共享物品因子的分解（图2）。然后，针对用户-物品关系矩阵中的数据是频次数据的特点，提出了一个基于泊松分布先验的多任务非负矩阵分解算法（算法1）。算法第一阶段，把用户-物品浏览矩阵和用户-物品购买矩阵一起分解为用户浏览兴趣因子矩阵和用户购买兴趣因子矩阵。第二阶段，把两个兴趣因子矩阵按列的最大值置1，其余为0，然后做两个因子矩阵的点积，生成所谓的兴趣转移字典。第三阶段，根据兴趣转移字典重构下一阶段用户-物品的购买矩阵，预测未来用户可能会购买何种物品（图3）。
结论：基于相同用户、相同物品的不同业务场景信息（如浏览行为数据和购买行为数据）分析用户潜在兴趣的概率产生模式，能够大幅度提升预测用户未来购买何种物品的效果。

关键词组：推荐系统；协同过滤；兴趣转移模型；跨域；隐性数据

Darkslateblue:Affiliate; Royal Blue:Author; Turquoise:Article

Reference

[1]Blei, D.M., Lafferty, J.D., 2006. Dynamic topic models. Int. Conf. on Machine Learning, p.113-120.

[2]Chen, G., Wang, F., Zhang, C., 2009. Collaborative filtering using orthogonal nonnegative matrix tri-factorization. Inform. Process. Manag., 45(3):368-379.

[3]Cohn, D., Hofmann, T., 2000. The missing link—a probabilistic model of document content and hypertext connectivity. Conf. on Neural Information Processing Systems, p.430-436.

[4]Devarajan, K., Wang, G., Ebrahimi, N., 2015. A unified statistical approach to non-negative matrix factorization and probabilistic latent semantic indexing. Mach. Learn., 99(1):137-163.

[5]Ding, C., Li, T., Peng, W., 2006. Nonnegative matrix factorization and probabilistic latent semantic indexing: equivalence, chi-square statistic, and a hybrid method. National Conf. on Artificial Intelligencer, p.342-347.

[6]Gaussier, E., Goutte, C., 2005. Relation between PLSA and NMF and implications. Int. ACM SIGIR Conf. on Research and Development in Information Retrieval, p.601-602.

[7]Gopalan, P., Hofman, J.M., Blei, D.M., 2013. Scalable recommendation with Poisson factorization. arXiv:1311.1704. http://arxiv.org/abs/1311.1704

[8]Gu, Q., Zhou, J., Ding, C., 2010. Collaborative filtering: weighted nonnegative matrix factorization incorporating user and item graphs. SIAM Int. Conf. on Data Mining, p.199-210.

[9]Herlocker, J.L., Konstan, J.A., Terveen, L.G., et al., 2004. Evaluating collaborative filtering recommender systems. ACM Trans. Inform. Syst., 22(1):5-53.

[10]Hofmann, T., 2004. Latent semantic models for collaborative filtering. ACM Trans. Inform. Syst., 22(1):89-115.

[11]Hofmann, T., Puzicha, J., 1999. Latent class models for collaborative filtering. Int. Joint Conf. on Artificial Intelligence, p.688-693.

[12]Ju, B., Qian, Y., Ye, M., et al., 2015. Using dynamic multi-task non-negative matrix factorization to detect the evolution of user preferences in collaborative filtering. PloS One, 10(8):e0135090.

[13]Koren, Y., 2010. Collaborative filtering with temporal dynamics. Commun. ACM, 53(4):89-97.

[14]Koren, Y., Bell, R., Volinsky, C., 2009. Matrix factorization techniques for recommender systems. Computer, 42(8):30-37.

[15]Lathia, N., Hailes, S., Capra, L., 2009. Temporal collaborative filtering with adaptive neighbourhoods. Int. ACM SIGIR Conf. on Research and Development in Information Retrieval, p.796-797.

[16]Lee, D.D., Seung, H.S., 1999. Learning the parts of objects by non-negative matrix factorization. Nature, 401(6755):788-791.

[17]Lee, D.D., Seung, H.S., 2000. Algorithms for non-negative matrix factorization. Conf. on Neural Information Processing Systems, p.556-562.

[18]Li, B., Yang, Q., Xue, X., 2009a. Can movies and books collaborate? Cross-domain collaborative filtering for sparsity reduction. Int. Joint Conf. on Artificial Intelligence, p.2052-2057.

[19]Li, B., Yang, Q., Xue, X., 2009b. Transfer learning for collaborative filtering via a rating-matrix generative model. Int. Conf. on Machine Learning, p.617-624.

[20]Li, T., Sindhwani, V., Ding, C., et al., 2010. Bridging domains with words: opinion analysis with matrix tri-factorizations. SIAM Int. Conf. on Data Mining, p.293-302.

[21]Liu, J., Wang, C., Gao, J., et al., 2013. Multi-view clustering via joint nonnegative matrix factorization. SIAM Int. Conf. on Data Mining, p.252-260.

[22]Ma, H., Liu, C., King, I., et al., 2011. Probabilistic factor models for web site recommendation. Int. ACM SIGIR Conf. on Research and Development in Information Retrieval, p.265-274.

[23]Mnih, A., Salakhutdinov, R., 2007. Probabilistic matrix factorization. Conf. on Neural Information Processing Systems, p.1257-1264.

[24]Rendle, S., Freudenthaler, C., 2014. Improving pairwise learning for item recommendation from implicit feedback. ACM Int. Conf. on Web Search and Data Mining, p.273-282.

[25]Rendle, S., Freudenthaler, C., Gantner, Z., et al., 2009. BPR: Bayesian personalized ranking from implicit feedback. 25th Conf. on Uncertainty in Artificial Intelligence, p.452-461.

[26]Salakhutdinov, R., Mnih, A., 2008. Bayesian probabilistic matrix factorization using Markov chain Monte Carlo. Int. Conf. on Machine Learning, p.880-887.

[27]Savia, E., Puolamäki, K., Kaski, S., 2009. Latent grouping models for user preference prediction. Mach. Learn., 74(1):75-109.

[28]Shi, Y., Karatzoglou, A., Baltrunas, L., et al., 2012. TFMAP: optimizing MAP for top-n context-aware recommendation. Int. ACM SIGIR Conf. on Research and Development in Information Retrieval, p.155-164.

[29]Si, L., Jin, R., 2003. Flexible mixture model for collaborative filtering. Int. Conf. on Machine Learning, p.704-711.

[30]Singh, A.P., Gordon, G.J., 2008. Relational learning via collective matrix factorization. 14th ACM SIGKDD Int. Conf. on Knowledge Discovery and Data Mining, p.650-658.

[31]Su, X., Khoshgoftaar, T.M., 2009. A survey of collaborative filtering techniques. Adv. Artif. Intell., 2009:421425.

[32]Xie, S., Lu, H., He, Y., 2012. Multi-task co-clustering via nonnegative matrix factorization. Int. Conf. on Pattern Recognition, p.2954-2958.

[33]Zhang, S., Wang, W., Ford, J., et al., 2006. Learning from incomplete ratings using non-negative matrix factorization. SIAM Int. Conf. on Data Mining, p.549-553.

[34]Zhuang, F., Luo, P., Xiong, H., et al., 2011. Exploiting associations between word clusters and document classes for cross-domain text categorization. Stat. Anal. Data Min.: ASA Data Sci. J., 4(1):100-114.

Open peer comments: Debate/Discuss/Question/Opinion

<1>

- Go to

基于兴趣转移模型的协同过滤算法

Darkslateblue:Affiliate; Royal Blue:Author; Turquoise:Article

Reference