Abstract
The Probabilistic Matrix Factorization (PMF) model has been widely studied for recommender systems, which outperform previous models with a solid probabilistic explanation. To further improve its accuracy by using social information, researchers attempt to combine the PMF model with social network graphs by adding social terms. However, existing works on social terms do not provide theoretical explanations to make the models well understood. The lack of explanations limits further improvement of prediction accuracy. Hence, in this paper we provide our explanation and propose a unified covariance framework to solve this problem. Our explanation, including regularization terms, factorization terms and an ensemble of them, reveals how most social terms work from a probabilistic view. Our framework shows that those terms could be optimized in a direct way compatible to PMF. We find out that accuracy improvements for existing works on regularization terms rely more on personalized properties, and that social information for factorization terms is helpful but not always necessary.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Chua, F.C.T., Lauw, H.W., Lim, E.-P.: Generative models for item adoptions using social correlation. IEEE Trans. Knowl. Data Eng. 25(9), 2036–2048 (2013)
Salakhutdinov, R., Mnih, A.: Probabilistic matrix factorization. In: Proceedings of the NIPS, pp. 1257–1264, 2007
Salakhutdinov, R. Mnih, A.: Bayesian probabilistic matrix factorization using Markov chain Monte Carlo. In: Proceedings of the ICML, pp. 880–887 (2008)
Jamali, M., Ester, M.: A matrix factorization technique with trust propagation for recommendation in social networks. In: Proceedings of the RecSys, pp. 135–142 (2010)
Ma, H., Zhou, D., Liu, C., Lyu, M.R., King, I.: Recommender systems with social regularization. In: Proceedings of the WSDM, pp. 287–296 (2011)
Ma, H.: An experimental study on implicit social recommendation. In: Proceedings of the SIGIR, pp. 73–82 (2013)
Yang, X., Steck, H., Liu, Y.: Circle-based recommendation in online social networks. In: Proceedings of the SIGKDD, pp. 1267–1275 (2012)
Forsati, R., Mahdavi, M., Shamsfard, M., Sarwat, M.: Matrix factorization with explicit trust and distrust side information for improved social recommendation. ACM Trans. Inform. Syst. 32(4) (2014)
Ma, H., Lyu, Ml.R., King, I.: Learning to recommend with trust and distrust relationships. In: Proceedings of the RecSys, pp. 189–196 (2009)
Fazeli, S., Loni, B., Bellogin, A., Drachsler, H., Sloep, P.: Implicit vs. explicit trust in social matrix factorization. In: Proceedings of the RecSys, pp. 317–320 (2014)
Xia, F., Liu, H., Asabere, N.Y., Wang, W., Yang, Z.: Multi-category item recommendation using neighborhood associations in trust networks. In: Proceedings of the WWW, pp. 403–404 (2014)
Ma, H., Yang, H., Lyu, Ml.R., King, I.: SoRec: social recommendation using probabilistic matrix factorization. In: Proceedings of the CIKM, pp. 931–940 (2008)
Yang, B., Yu, L., Liu, D., Liu, J.: Social collaborative filtering by trust. In: Proceedings of the IJCAI, pp. 2747–2753 (2013)
Zhang, Y., Lai, G., Zhang, M., Zhang, Y., Liu, Y., Ma, S.: Explicit factor models for explainable recommendation based on phrase-level sentiment analysis. In: Proceedings of the SIGIR, pp. 83–92 (2014)
Jiang, M., Cui, P., Liu, R., Yang, Q., Wang, F., Zhu, W., Yang, S.: Social contextual recommendation. In: Proceedings of the CIKM, pp. 45–54 (2012)
Wang, T., Jin, X., Ding, X., Ye, X.: User interests imbalance exploration in social recommendation: a fitness adaptation. In: Proceedings of the CIKM, pp. 281–290 (2014)
Jiang, Y., Liu, J., Zhang, X., Li, Z., Lu. H.: TCRec: product recommendation via exploiting social-trust network and product category information. In: Proceedings of the WWW, pp. 233–234 (2013)
Yao, W., He, J., Huang, G., Zhang, Y.: Modeling dual role preferences for trust-aware recommendation. In: Proceedings of the SIGIR, pp. 975–978 (2014)
Zhao, T., Li, C., Li, M., Ding, Q., Li, L.: Social recommendation incorporating topic mining and social trust analysis. In: Proceedings of the CIKM, pp. 1643–1648 (2013)
Yuan, Q., Chen, L., Zhao, S.: Factorization vs. regularization: integrating heterogeneous social relationships in top-n recommendation. In: Proceedings of the RecSys, pp. 245–252 (2011)
Guo, L., Ma, J., Chen, Z.: Learning to recommend with multi-faceted trust in social networks. In: Proceedings of the WWW, pp. 205–206 (2013)
Feng, H., Qian, X.: Recommendation via user’s personality and social contextual. In: Proceedings of the CIKM, pp. 1521–1524 (2013)
Tang, J., Hu, X., Gao, H., Liu, H.: Exploiting local and global social context for recommendation. In: Proceedings of the AAAI, pp. 2712–2718 (2013)
Acknowledgments
This work is partially supported by National Nature Science Foundation of China (No. 61374199, National High-tech R&D Program (No. 2015AA042101),) and Beijing Natural Science Foundation (No. 4142031).
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2017 Springer International Publishing AG
About this paper
Cite this paper
Liu, J., Zhang, L., Sheng, V.S., Laili, Y. (2017). Social Recommendation Terms: Probabilistic Explanation Optimization. In: Zhang, L., Ren, L., Kordon, F. (eds) Challenges and Opportunity with Big Data. Monterey Workshop 2016. Lecture Notes in Computer Science(), vol 10228. Springer, Cham. https://doi.org/10.1007/978-3-319-61994-1_15
Download citation
DOI: https://doi.org/10.1007/978-3-319-61994-1_15
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-61993-4
Online ISBN: 978-3-319-61994-1
eBook Packages: Computer ScienceComputer Science (R0)