ABSTRACT
Data sparsity is one of the biggest problems faced by collaborative filtering used in recommender systems. Data imputation alleviates the data sparsity problem by inferring missing ratings and imputing them to the original rating matrix. In this paper, we identify the limitations of existing data imputation approaches and suggest three new claims that all data imputation approaches should follow to achieve high recommendation accuracy. Furthermore, we propose a deep-learning based approach to compute imputed values that satisfies all three claims. Based on our hypothesis that most pre-use preferences (e.g., impressions) on items lead to their post-use preferences (e.g., ratings), our approach tries to understand via deep learning how pre-use preferences lead to post-use preferences differently depending on the characteristics of users and items. Through extensive experiments on real-world datasets, we verify our three claims and hypothesis, and also demonstrate that our approach significantly outperforms existing state-of-the-art approaches.
- G. Adomavicius and A. Tuzhilin. 2005. Toward the Next Generation of Recommender Systems: A Survey of the State-of-the-Art and Possible Extensions. IEEE Trans. on Knowledge and Data Engineering Vol. 17 (2005), 734--749. Google ScholarDigital Library
- S. Chang, F. M. Harper, and L. G. Terveen. 2016. Crowd-Based Personalized Natural Language Explanations for Recommendations Proc. ACM Int'l Conf. on Recommender Systems. 175--182. Google ScholarDigital Library
- P. Cremonesi, Y. Koren, and R. Turrin. 2010. Performance of Recommender Algorithms on Top-N Recommendation Tasks Proc. ACM Int'l Conf. on Recommender Systems. 39--46. Google ScholarDigital Library
- C. Doersch. 2016. Tutorial on Variational Autoencoders. arXiv preprint arXiv:1606.05908 (2016).Google Scholar
- B. Sarwar et al.. 2001. Item-Based Collaborative Filtering Recommendation Algorithms Proc. Int'l Conf. on World Wide Web. 285--295. Google ScholarDigital Library
- D. Kingma et al.. 2014 a. Semi-supervised Learning with Deep Generative Models Advances in Neural Information Processing Systems. 3581--3589. Google ScholarDigital Library
- D. K. Chae et al.. 2018 a. On Identifying k-Nearest Neighbors in Neighborhood Models for Efficient and Effective Collaborative Filtering. Neurocomputing Vol. 278 (2018), 124--143.Google ScholarCross Ref
- J. Lee et al.. 2013 a. Alleviating the Sparsity in Collaborative Filtering using Crowdsourcing Proc. of Workshop on Crowdsourcing and Human Computation for Recommender Systems.Google Scholar
- J. Lee et al.. 2016 a. Improving the Accuracy of Top-N Recommendation using a Preference Model. Information Sciences Vol. 348 (2016), 290--304. Google ScholarDigital Library
- J. W. Ha et al.. 2012 a. Top-N Recommendation through Belief Propagation. In Proc. ACM Int'l Conf. on Information and Knowledge Management. 2343--2346. Google ScholarDigital Library
- M. H. Jang et al.. 2016 b. PIN-TRUST: Fast Trust Propagation Exploiting Positive, Implicit, and Negative Information. In Proc. ACM Int'l Conf. on Information and Knowledge Management. 629--638. Google ScholarDigital Library
- R. Pan et al.. 2008. One-Class Collaborative Filtering. In Proc. IEEE Int'l Conf. on Data Mining. 502--511. Google ScholarDigital Library
- S. C. Lee et al.. 2015 a. A Graph-Based Recommendation Framework for Price-Comparison Services Proc. Int'l Conf. on World Wide Web. 59--60. Google ScholarDigital Library
- S. C. Lee et al.. 2017. A Single-Step Approach to Recommendation Diversification Proc. Int'l Conf. on World Wide Web. 809--810. Google ScholarDigital Library
- S. Rendle et al.. 2009. BPR: Bayesian Personalized Ranking from Implicit Feedback Proc. Int'l Conf. on Uncertainty in Artificial Intelligence. 452--461. Google ScholarDigital Library
- V. Sindhwani et al.. 2010 a. One-Class Matrix Completion with Low-Density Factorizations Proc. IEEE Int'l Conf. on Data Mining. 1055--1060. Google ScholarDigital Library
- W. S. Hwang et al.. 2013 b. Exploiting Trustors as well as Trustees in Trust-Based Recommendation Proc. ACM Int'l Conf. on Information and Knowledge Management. 1893--1896. Google ScholarDigital Library
- W. S. Hwang et al.. 2014 b. Data Imputation using a Trust Network for Recommendation Proc. Int'l Conf. on World Wide Web. 299--300. Google ScholarDigital Library
- W. S. Hwang et al.. 2015 b. On Exploiting Trustors in Trust-Based Recommendation. Journal of Internet Technology Vol. 16, 4 (2015), 755--765.Google Scholar
- W. S. Hwang et al.. 2016 c. Efficient Recommendation Methods using Category Experts for a Large Dataset. Information Fusion Vol. 28 (2016), 75--82. Google ScholarDigital Library
- W. S. Hwang et al.. 2016 d. “Told You I Didn't Like It”: Exploiting Uninteresting Items for Effective Collaborative Filtering. In Proc. IEEE Int'l Conf. on Data Engineering. 349--360.Google ScholarCross Ref
- X. He et al.. 2016 e. Fast Matrix Factorization for Online Recommendation with Implicit Feedback Proc. ACM Int'l Conf. on Research and Development in Information Retrieval. 549--558. Google ScholarDigital Library
- Y. C. Lee et al.. 2018 b. gOCCF: Graph-Theoretic One-Class Collaborative Filtering Based on Uninteresting Items Proc. AAAI Conf. on Artificial Intelligence.Google Scholar
- Y. Li et al.. 2010 b. Improving One-Class Collaborative Filtering by Incorporating rich User Information Proc. ACM Int'l Conf. on Information and Knowledge Management. 959--968. Google ScholarDigital Library
- Y. Pu et al.. 2016 f. Variational Autoencoder for Deep Learning of Images, Labels and Captions Advances in Neural Information Processing Systems. 2352--2360. Google ScholarDigital Library
- Y. Ren et al.. 2012 b. The Efficient Imputation Method for Neighborhood-Based Collaborative Filtering. In Proc. ACM Int'l Conf. on Information and Knowledge Management. 684--693. Google ScholarDigital Library
- Y. Ren et al.. 2013 c. AdaM: Adaptive-Maximum Imputation for Neighborhood-Based Collaborative Filtering. In Proc. IEEE Int'l Conf. on Advances in Social Networks Analysis and Mining. 628--635. Google ScholarDigital Library
- Y. Yao et al.. 2014 c. Dual-Regularized One-Class Collaborative Filtering Proc. ACM Int'l Conf. on Information and Knowledge Management. 759--768. Google ScholarDigital Library
- Z. Gantner et al.. 2011. MyMediaLite: A Free Recommender System Library. In Proc. ACM Int'l Conf. on Recommender Systems. 305--308. Google ScholarDigital Library
- R. He and J. McAuley. 2016. Ups and Downs: Modeling the Visual Evolution of Fashion Trends with One-Class Collaborative Filtering. In Proc. Int'l Conf. on World Wide Web. 507--517. Google ScholarDigital Library
- Y. Hu, Y. Koren, and C. Volinsky. 2008. Collaborative Filtering for Implicit Feedback Datasets Proc. IEEE Int'l Conf. on Data Mining. 263--272. Google ScholarDigital Library
- D. Kingma and M. Welling. 2013. Auto-Encoding Variational Bayes. In arXiv preprint arXiv:1312.6114.Google Scholar
- Y. Koren. 2008. Factorization Meets the Neighborhood: A Multifaceted Collaborative Filtering Model. In Proc. ACM Int'l Conf. on Knowledge Discovery and Data Mining. 426--434. Google ScholarDigital Library
- Y. Koren, R. Bell, and C. Volinsky. 2009. Matrix Factorization Techniques for Recommender Systems. Computer Vol. 42 (2009), 30--37. Google ScholarDigital Library
- X. Li and J. She. 2017. Collaborative Variational Autoencoder for Recommender Systems Proc. of ACM Int'l Conf. on Knowledge Discovery and Data Mining. 305--314. Google ScholarDigital Library
- H. Ma, I. King, and M. R. Lyu. 2007. Effective Missing Data Prediction for Collaborative Filtering Proc. ACM Int'l Conf. on Research and Development in Information Retrieval. 39--46. Google ScholarDigital Library
- S. Nathan and J. Tommi. 2003. Weighted Low-Rank Approximations. In Proc. Int'l Conf. on Machine Learning. 720--727. Google ScholarDigital Library
- R. Pan and M. Scholz. 2009. Mind the Gaps: Weighting the Unknown in Large-Scale One-Class Collaborative Filtering Proc. ACM Int'l Conf. on Knowledge Discovery and Data Mining. 667--676. Google ScholarDigital Library
- W. Pan and L. Chen. 2013. GBPR: Group Preference Based Bayesian Personalized Ranking for One-Class Collaborative Filtering. In Proc. Int'l Joint Conf. on Artificial Intelligence. 2691--2697. Google ScholarDigital Library
- S. Rendle and C. Freudenthaler. 2014. Improving Pairwise Learning for Item Recommendation from Implicit Feedback Proc. Int'l Conf. on Web Search and Data Mining. 273--282. Google ScholarDigital Library
- R. Salakhutdinov and A. Mnih. 2008. Probabilistic Matrix Factorization. In Advances in Neural Information Processing Systems. 1257--1264. Google ScholarDigital Library
- B. Smyth, R. Rafter, and S. Banks. 2016. Harnessing Crowdsourced Recommendation Preference Data from Casual Gameplay Proc. ACM Int'l Conf. on User Modeling Adaptation and Personalization. 95--104. Google ScholarDigital Library
- H. Steck. 2010. Training and Testing of Recommender Systems on Data Missing Not at Random Proc. ACM Int'l Conf. on Knowledge Discovery and Data Mining. 713--722. Google ScholarDigital Library
- M. Volkovs and G. W. Yu. 2015. Effective Latent Models for Binary Feedback in Recommender Systems Proc. ACM Int'l Conf. on Research and Development in Information Retrieval. 313--322. Google ScholarDigital Library
Index Terms
- How to Impute Missing Ratings?: Claims, Solution, and Its Application to Collaborative Filtering
Recommendations
Merging trust in collaborative filtering to alleviate data sparsity and cold start
Providing high quality recommendations is important for e-commerce systems to assist users in making effective selection decisions from a plethora of choices. Collaborative filtering is a widely accepted technique to generate recommendations based on ...
An effective recommender system by unifying user and item trust information for B2B applications
Although Collaborative Filtering (CF)-based recommender systems have received great success in a variety of applications, they still under-perform and are unable to provide accurate recommendations when users and items have few ratings, resulting in ...
A Hybrid Multigroup Coclustering Recommendation Framework Based on Information Fusion
Special Section on Visual Understanding with RGB-D SensorsCollaborative Filtering (CF) is one of the most successful algorithms in recommender systems. However, it suffers from data sparsity and scalability problems. Although many clustering techniques have been incorporated to alleviate these two problems, ...
Comments