ABSTRACT
The implicit feedback based recommendation problem---when only the user history is available but there are no ratings---is a much harder task than the explicit feedback based recommendation problem, due to the inherent uncertainty of the interpretation of such user feedbacks. Still, this practically important recommendation task received less attention and therefore there are only a few common implicit feedback based algorithms and benchmark datasets. This paper focuses on a common matrix factorization method for the implicit problem and investigates if recommendation performance can be improved by appropriate initialization of the feature vectors before training. We present a general initialization framework that preserves the similarity between entities (users/items) when creating the initial feature vectors, where similarity is defined using e.g. context or metadata information. We demonstrate how the proposed initialization framework can be coupled with MF algorithms. The efficiency of the initialization is evaluated using various context and metadata based similarity concepts on two implicit variants of the MovieLens 10M dataset and one real life implicit database. It is shown that performance gain can attain 10% improvement in recall@50 and in AUC@50.
- Bell, R. M., and Koren, Y. Scalable collaborative filtering with jointly derived neighborhood interpolation weights. In Proc of. ICDM-07, 7th IEEE Int. Conf. on Data Mining (Omaha, Nebraska, USA, 2007), 43--52. Google ScholarDigital Library
- Bennett, J., and Lanning, S. The Netflix Prize. In Proc. of KDD Cup Workshop at SIGKDD-07, 13th ACM Int. Conf. on Knowledge Discovery and Data Mining (San Jose, California, USA, 2007), 3--6.Google Scholar
- Boutsidis, C., and Gallopoulos, E. Svd based initialization: A head start for nonnegative matrix factorization, 2007.Google Scholar
- Dhillon, I. S., and Modha, D. S. Concept decompositions for large sparse text data using clustering. In Machine Learning (2000), 143--175. Google ScholarDigital Library
- Herlocker, J. L., Konstan, J. A., Terveen, L. G., and Riedl, J. T. Evaluating collaborative filtering recommender systems. ACM Transactions on Information Systems 22, 1 (2004), 5--53. Google ScholarDigital Library
- Hu, Y., Koren, Y., and Volinsky, C. Collaborative filtering for implicit feedback datasets. In Proc. of ICDM-08, 8th IEEE Int. Conf. on Data Mining (Pisa, Italy, December 15--19, 2008), 263--272. Google ScholarDigital Library
- Jolliffe, I. Principal Component Analysis. Springer Verlag, 1986.Google ScholarCross Ref
- Koren, Y. Factorization meets the neighborhood: a multifaceted collaborative filtering model. In Proc. of the 14th ACM Int. Conf. on Knowledge Discovery and Data Mining (SIGKDD'08) (Las Vegas, Nevada, USA, 2008), 426--434. Google ScholarDigital Library
- Kramer, M. A. Nonlinear principal component analysis using autoassociative neural networks. AIChE Journal 37, 2 (1991), 233--243.Google ScholarCross Ref
- Langville, A. N., Meyer, C. D., and Albright, R. Initializations for the nonnegative matrix factorization, 2006.Google Scholar
- Pilászy, I., and Tikk, D. Recommending new movies: Even a few ratings are more valuable than metadata. In Proc. of the 3rd ACM Conf. on Recommender Systems (Recsys'09), ACM (New York, NY, USA, 2009), 93--100. Google ScholarDigital Library
- Pilászy, I., Zibriczky, D., and Tikk, D. Fast ALS-based matrix factorization for explicit and implicit feedback datasets. In Proc. of the 4th ACM Conf. on Recommender Systems (RecSys'10), ACM (Barcelona, Spain, 2010), 71--78. Google ScholarDigital Library
- Rendle, S., and Schmidt-Thieme, L. Online-updating regularized kernel matrix factorization models for large-scale recommender systems. In Proc. of RecSys-08: ACM Conf. on Recommender Systems, ACM (New York, NY, USA, 2008), 251--258. Google ScholarDigital Library
- Resnick, P., Iacovou, N., Suchak, M., Bergstrom, P., and Riedl, J. Grouplens: an open architecture for collaborative filtering of netnews. In Proc. of CSCW-94, 4th ACM Conf. on Computer Supported Cooperative Work (Chapel Hill, North Carolina, USA, 1994), 175--186. Google ScholarDigital Library
- Ricci, F., Rokach, L., and Shapira, B. Introduction to recommender systems handbook. In Recommender Systems Handbook, F. Ricci, L. Rokach, B. Shapira, and P. B. Kantor, Eds., Artificial Intelligence. Springer US, 2011, 1--35.Google Scholar
- Salakhutdinov, R., and Mnih, A. Probabilistic matrix factorization. In Advances in Neural Information Processing Systems 20, J. C. Platt, D. Koller, Y. Singer, and S. Roweis, Eds. MIT Press, Cambridge, Massachusetts, USA, 2008.Google ScholarDigital Library
- Smilde, A., Bro, R., and Geladi, P. Multi-way Analysis. Wiley, West Sussex, England, 2004.Google Scholar
- Snedecor, G. W., and Cochran, W. G. Statistical Methods, 7th ed. Iowa State University Press, 1980.Google Scholar
- Takács, G., Pilászy, I., Németh, B., and Tikk, D. Major components of the gravity recommendation system. SIGKDD Explor. Newsl. 9 (December 2007), 80--83. Google ScholarDigital Library
- Takács, G., Pilászy, I., Németh, B., and Tikk, D. Scalable collaborative filtering approaches for large recommender systems. Journal of Machine Learning Research 10 (2009), 623--656. Google ScholarDigital Library
- Takács, G., Pilászy, I., and Tikk, D. Applications of the conjugate gradient method for implicit feedback collaborative filtering. In RecSys'11: Proc. of the 4th ACM Conf. on Recommender Systems (Chicago, IL, USA, October 23--27, 2011), 297--300. Google ScholarDigital Library
- Thai-Nghe, N., Drumond, L., Horváth, T., Krohn-Grimberghe, A., Nanopoulos, A., and Schmidt-Thieme, L. Factorization techniques for predicting student performance. In Educational Recommender Systems and Technologies: Practices and Challenges (ERSAT 2011). IGI Global, 2012, 1--25.Google Scholar
- Wild, S., Wild, W. S., Curry, J., Dougherty, A., and Betterton, M. Seeding non-negative matrix factorization with the spherical k-means clustering. Tech. rep., 2003.Google Scholar
- Wild, S. M., Curry, J. H., and Dougherty, A. Improving non-negative matrix factorizations through structured initialization. Pattern Recognition 37, 11 (November 2004), 2217--2232.Google ScholarDigital Library
Index Terms
- Enhancing matrix factorization through initialization for implicit feedback databases
Recommendations
Incremental Matrix Co-factorization for Recommender Systems with Implicit Feedback
WWW '18: Companion Proceedings of the The Web Conference 2018Recommender systems try to predict which items a user will prefer. Traditional models for recommendation only take into account the user-item interaction, usually expressed by explicit ratings. However, in these days, web services continuously generate ...
Using Implicit Feedback for Neighbors Selection: Alleviating the Sparsity Problem in Collaborative Recommendation Systems
WebMedia '17: Proceedings of the 23rd Brazillian Symposium on Multimedia and the WebThe most popular Recommender systems (RSs) employ Collaborative Filtering (CF) algorithms where users explicitly rate items. Based on these ratings, a user-item rating matrix is generated and used to select the items to be recommended for a target user. ...
A latent pairwise preference learning approach for recommendation from implicit feedback
CIKM '12: Proceedings of the 21st ACM international conference on Information and knowledge managementMost of the current recommender systems heavily rely on explicit user feedback such as ratings on items to model users' interests. However, in many applications, it is very hard to collect the explicit feedback, while implicit feedback such as user ...
Comments