In this paper we proposed a novel method to allevi- ACM SIGIR conference on Research ate the sparsity problem and improve the quality of the development in informaion retrieval, collaborative filtering method. We make use of the tag 259-266. New York. NY USA. 2003 information to find closer neighbors for the users and the items, respectively. These neighbors give stron 7. T Hofmann. Latent semantic models for inferences for the potential preference of the user for collaborative filtering. ACM Trans. Inf. Syst the item, We utilize these inferences to make the rat- 22(1):89-115,2004. ing prediction. According to the experiments, our ap 8. Y. Koren. Factorization meets the neighborhood proach's prediction for the users' preference is much more accurate than the art-of-state ones such as nmf a multifaceted collaborative filtering model. In method, PMF method and the improved regularized KDD, pages426-434,2008 SVD method 9. W. Li and D. Yeung. Relation regularized matrix factorization. In Proceedings of the 21 st Finding neighbors is vital for an excellent collabora International Joint Conference on Artificial tive filtering algorithm. The motivation of our work Intellig 2009 is to find better neighbors, which give stronger infer ence for the prediction. Latent topics connect the users .O. H. Ma, H. Yang, M.R. Lyu, and I. King. Sorec: and items with similar interests together. Finding these Social recommendation sing probabilistic matrix topics is equal to finding the neighbors. The connection factorization In CIKM, pages 931-940, 2008 that cannot be discovered in the rating records can be disclosed through learning about the tagging histor 11. A. Marchetti. M. Tesconi. F. Ronzano. M. rosella This is why our method outperforms the others and S. Minutoli. SemKey: A semantic collaborative tagging system. In Proceedings of The next step for us is to improve our method in two 16th International World wide web Conference, vww2007 Citeseer 2007 aspects. One is incrementalization. The neighborho nethod enjoys the low computational complexity but 12. B. Marlin. Collaborative filtering: A machine suffers from the rigidity to frequent update. If there are learning perspective. Master's thesis, University of lot of new entries from the users to items our current Toronto. 2004 solution fail to deal with this situation on the other an fuse the collaborative matrix factorization 13. P. Massa and p avesani. Trust-aware nethod with our topic finding model. This is another recommender systems. In RecSys 07: Proceedings way to make use of the latent topic information. We of the 2007 ACM conference on Recommender believe these methods are both promising solutions to systems, pages 17-24, New York, NY, USA, 2007. further improve the collaborative filtering technique 14. B J. Mirza, B. J. Keller, and N. Ramakrishnan. ADDITIONAL AUTHORS Studying recommendation algorithms by graph REFERENCES analysis. Journal of Intelligent Information 1. R. Bell and Y. Koren. Improved Systems,20(2):131-160, March2003 neighborhood-based collaborative filtering. In KDD-Cup and Workshop. Citeseer, 2007 15. N. S. Nati and T Jaakkola. Weighted low-rank approximations. InIn 20th International 2. R. M. Bell and Y. Koren. Scalable collaborative Conference on Machine Learning, pages 720-727 filtering with jointly derived neighborhood aaai Press. 2003 terpolation weights In ICDM, pages 43-52 16. R. Salakhutdinov and A. mnih. Probabilistic matrix factorization. In Advances in Neural 3. D. M. Blei, A. Y. Ng, and M. I Jordan. Latent Information Processing Systems, 2007 dirichlet allocation. Journal of Machine Learning Research,3:993-1022,2003 17. Connecting users to items through tags. In ww S. Sen,J. Vig, and J. Riedl. Tagommenders 4. D. Goldberg, D. Nichols, B. M. Oki, and D. Terry. t sing collaborative filtering to weave an formation tapestry. Commun. ACM, 18. B. Sigurbj 35(12):61-70,1992 ornsson and R. Van Zwol. Flickr tag 5. D. M. Hawkins, S. C. Basak, and D. Mills Assessing model fit by cross-validation. J. Chem Inf. Comput. Sci., 43(2): 579-586, March 2003 19. A. Singh and G. Gordon. Relational learning via collective matrix factorization In Proceeding of the 6. T. Hofmann. Collaborative filtering via gaussian 4th ACM SIGKDD international conference on probabilistic latent semantic analysis. In SIGIR Knowledge discovery and data 03: Proceedings of the 26th annual international 650-658.ACM,2008In this paper we proposed a novel method to allevi￾ate the sparsity problem and improve the quality of the collaborative filtering method. We make use of the tag information to find closer neighbors for the users and the items, respectively. These neighbors give strong inferences for the potential preference of the user for the item. We utilize these inferences to make the rat￾ing prediction. According to the experiments, our ap￾proach’s prediction for the users’ preference is much more accurate than the art-of-state ones such as NMF method, PMF method and the improved regularized SVD method. Finding neighbors is vital for an excellent collabora￾tive filtering algorithm. The motivation of our work is to find better neighbors, which give stronger infer￾ence for the prediction. Latent topics connect the users and items with similar interests together. Finding these topics is equal to finding the neighbors. The connection that cannot be discovered in the rating records can be disclosed through learning about the tagging history. This is why our method outperforms the others. The next step for us is to improve our method in two aspects. One is incrementalization. The neighborhood method enjoys the low computational complexity, but suffers from the rigidity to frequent update. If there are a lot of new entries from the users to items, our current solution fail to deal with this situation. On the other hand, we can fuse the collaborative matrix factorization method with our topic finding model. This is another way to make use of the latent topic information. We believe these methods are both promising solutions to further improve the collaborative filtering technique. ADDITIONAL AUTHORS REFERENCES 1. R. Bell and Y. Koren. Improved neighborhood-based collaborative filtering. In KDD-Cup and Workshop. Citeseer, 2007. 2. R. M. Bell and Y. Koren. Scalable collaborative filtering with jointly derived neighborhood interpolation weights. In ICDM, pages 43–52, 2007. 3. D. M. Blei, A. Y. Ng, and M. I. Jordan. Latent dirichlet allocation. Journal of Machine Learning Research, 3:993–1022, 2003. 4. D. Goldberg, D. Nichols, B. M. Oki, and D. Terry. Using collaborative filtering to weave an information tapestry. Commun. ACM, 35(12):61–70, 1992. 5. D. M. Hawkins, S. C. Basak, and D. Mills. Assessing model fit by cross-validation. J. Chem. Inf. Comput. Sci., 43(2):579–586, March 2003. 6. T. Hofmann. Collaborative filtering via gaussian probabilistic latent semantic analysis. In SIGIR ’03: Proceedings of the 26th annual international ACM SIGIR conference on Research and development in informaion retrieval, pages 259–266, New York, NY, USA, 2003. ACM. 7. T. Hofmann. Latent semantic models for collaborative filtering. ACM Trans. Inf. Syst., 22(1):89–115, 2004. 8. Y. Koren. Factorization meets the neighborhood: a multifaceted collaborative filtering model. In KDD, pages 426–434, 2008. 9. W. Li and D. Yeung. Relation regularized matrix factorization. In Proceedings of the 21st International Joint Conference on Artificial Intelligence, 2009. 10. H. Ma, H. Yang, M. R. Lyu, and I. King. Sorec: Social recommendation using probabilistic matrix factorization. In CIKM, pages 931–940, 2008. 11. A. Marchetti, M. Tesconi, F. Ronzano, M. Rosella, and S. Minutoli. SemKey: A semantic collaborative tagging system. In Proceedings of 16th International World Wide Web Conference, WWW2007. Citeseer, 2007. 12. B. Marlin. Collaborative filtering: A machine learning perspective. Master’s thesis, University of Toronto, 2004. 13. P. Massa and P. Avesani. Trust-aware recommender systems. In RecSys ’07: Proceedings of the 2007 ACM conference on Recommender systems, pages 17–24, New York, NY, USA, 2007. ACM. 14. B. J. Mirza, B. J. Keller, and N. Ramakrishnan. Studying recommendation algorithms by graph analysis. Journal of Intelligent Information Systems, 20(2):131–160, March 2003. 15. N. S. Nati and T. Jaakkola. Weighted low-rank approximations. In In 20th International Conference on Machine Learning, pages 720–727. AAAI Press, 2003. 16. R. Salakhutdinov and A. Mnih. Probabilistic matrix factorization. In Advances in Neural Information Processing Systems, 2007. 17. S. Sen, J. Vig, and J. Riedl. Tagommenders: Connecting users to items through tags. In WWW, 2009. 18. B. Sigurbj ”ornsson and R. Van Zwol. Flickr tag recommendation based on collective knowledge. 2008. 19. A. Singh and G. Gordon. Relational learning via collective matrix factorization. In Proceeding of the 14th ACM SIGKDD international conference on Knowledge discovery and data mining, pages 650–658. ACM, 2008