Personalized topic modeling for recommending user-generated content

Abstract

User-generated content (UGC) such as blogs and twitters are exploding in modern Internet services. In such systems, recommender systems are needed to help people filter vast amount of UGC generated by other users. However, traditional rec-ommendation models do not use user authorship of items. In this paper, we show that with this additional information, we can significantly improve the performance of recommendations. A generative model that combines hierarchical topic modeling and matrix factorization is proposed. Empirical results show that our model outperforms other state-of-the-art models, and can provide interpretable topic structures for users and items. Furthermore, since user interests can be inferred from their productions, rec-ommendations can be made for users that do not have any ratings to solve the cold-start problem.

Keywords

User-generated content (UGC) / Collaborative filtering (CF) / Matrix factorization (MF) / Hierarchical topic modeling

References

[1] Agarwal,D., Chen,B.C., 2009. Regression-based latent factor models. Proc. 15th ACM SIGKDD Int. Conf. on Knowledge Discovery and Data Mining, p.19–28. http://dx.doi.org/10.1145/1557019.1557029

[2] Agarwal,D., Chen,B.C., 2010. fLDA: matrix factorization through latent Dirichlet allocation. Proc. 3rd ACM Int. Conf. on Web Search and Data Mining, p.91–100. http://dx.doi.org/10.1145/1718487.1718499

[3] Blei,D.M., Ng,A.Y., Jordan,M.I. , 2003. Latent Dirichlet allocation. J Mach. Learn. Res., 3:993–1022.

[4] Blei,D.M., Griffiths, T.L., Jordan,M.I. , 2010. The nested Chinese restaurant process and Bayesian nonparametric inference of topic hierarchies. J. ACM, 57(2):7. http://dx.doi.org/10.1145/1667053.1667056

[5] de Pessemier,T., Deryckere, T., Martens,L. , 2011. Context aware recommendations for user-generated content on a social network site. Proc. 7th European Interactive Tele-vision Conf., p.133–136. http://dx.doi.org/10.1145/1542084.1542108

[6] Deshpande,M., Karypis, G., 2004. Item-based top-N recom-mendation algorithms. ACM Trans. Inform. Syst., 22(1): 143–177. http://dx.doi.org/10.1145/963770.963776

[7] Duchi,J., Shalev-Shwartz, S., Singer,Y. , , 2008. Efficient projections onto the ℓ1-ball for learning in high dimen-sions. Proc. 25th Int. Conf. on Machine Learning, p.272–279.

[8] Hu,Y., Koren, Y., Volinsky,C. , 2008. Collaborative filtering for implicit feedback datasets. 8th IEEE Int. Conf. on Data Mining, p.263–272. http://dx.doi.org/10.1109/ICDM.2008.22

[9] Li,Y.M., Yang,M., Zhang,Z.F. , 2013. Scientific articles recommendation. Proc. 22nd ACM Int. Conf. on Infor-mation and Knowledge Management, p.1147–1156. http://dx.doi.org/10.1145/2505515.2505705

[10] Linden,G., Smith, B., York,J. , 2003. Amazon.com recom-mendations: item-to-item collaborative filtering. IEEE Intern. Comput., 7(1):76–80. http://dx.doi.org/10.1109/MIC.2003.1167344

[11] Lops,P., de Gemmis, M., Semeraro,G. , 2011. Content-based recommender systems: state of the art and trends. In: Ricci, F., Rokach, L., Shapira, B., (Eds.), Recom-mender Systems Handbook. Springer, Boston, p.73–105. http://dx.doi.org/10.1007/978-0-387-85820-3_3

[12] Melville,P., Mooney, R.J., Nagarajan,R. , 2002. Content- boosted collaborative filtering for improved recommen-dations. Proc. 8th National Conf. on Artificial Intelli-gence, p.187–192.

[13] Mooney,R.J., Roy,L., 2000. Content-based book recom-mending using learning for text categorization. Proc. 5th ACM Conf. on Digital Libraries, p.195–204. http://dx.doi.org/10.1145/336597.336662

[14] Pan,R., Zhou,Y., Cao,B., , 2008. One-class collaborative filtering. 8th IEEE Int. Conf. on Data Mining, p.502–511. http://dx.doi.org/10.1109/ICDM.2008.16

[15] Purushotham,S., Liu, Y., Kuo,C.C.J. , 2012. Collaborative topic regression with social matrix factorization for recommendation systems. arXiv:1206.4684.

[16] Rendle,S., Freudenthaler, C., Gantner,Z. , , 2009. BPR: Bayesian personalized ranking from implicit feedback. Proc. 25th Conference on Uncertainty in Artificial Intel-ligence, p.452–461.

[17] Salakhutdinov,R., Mnih, A., 2007. Probabilistic matrix fac-torization. Neural Information Processing Systems, p.1257–1264.

[18] Salakhutdinov,R., Mnih, A., 2008. Bayesian probabilistic matrix factorization using Markov chain Monte Carlo. Proc. 25th Int. Conf. on Machine Learning, p.880–887. http://dx.doi.org/10.1145/1390156.1390267

[19] Teh,Y.W., Jordan, M.I., Beal,M.J. , , 2004. Sharing clusters among related groups: hierarchical Dirichlet processes. Neural Information Processing Systems, p.1385–1392.

[20] Veeramachaneni,S., Sona, D., Avesani,P. , 2005. Hierarchical Dirichlet model for document classification. Proc. 22nd Int. Conf. on Machine Learning, p.928–935. http://dx.doi.org/10.1145/1102351.1102468

[21] Wang,C., Blei,D.M., 2011. Collaborative topic modeling for recommending scientific articles. Proc. 17th ACM SIGKDD Int. Conf. on Knowledge Discovery and Data Mining, p.448–456. http://dx.doi.org/10.1145/2020408.2020480

[22] Xu,Y., Yin,J., 2015. Collaborative recommendation with user generated content. Eng. Appl. Artif. Intel., 45(C):281–294. http://dx.doi.org/10.1016/j.engappai.2015.07.012

[23] Xu,Y., Chen,Z., Yin,J., , 2015. Learning to recommend with user generated content. Int. Conf. on Web-Age In-formation Management, p.221–232. http://dx.doi.org/10.1007/978-3-319-21042-1_18

RIGHTS & PERMISSIONS

2017 Zhejiang University and Springer-Verlag Berlin Heidelberg