ABSTRACT
We describe an innovative and scalable recommendation system successfully deployed at eBay. To build recommenders for long-tail marketplaces requires projection of volatile items into a persistent space of latent products. We first present a generative clustering model for collections of unstructured, heterogeneous, and ephemeral item data, under the assumption that items are generated from latent products. An item is represented as a vector of independently and distinctly distributed variables, while a latent product is characterized as a vector of probability distributions, respectively. The probability distributions are chosen as natural stochastic models for different types of data. The learning objective is to maximize the total intra-cluster coherence measured by the sum of log likelihoods of items under such a generative process. In the space of latent products, robust recommendations can then be derived using naive Bayes for ranking, from historical transactional data. Item-based recommendations are achieved by inferring latent products from unseen items. In particular, we develop a probabilistic scoring function of recommended items, which takes into account item-product membership, product purchase probability, and the important auction-end-time factor. With the holistic probabilistic measure of a prospective item purchase, one can further maximize the expected revenue and the more subjective user satisfaction as well. We evaluated the latent product clustering and recommendation ranking models using real-world e-commerce data from eBay, in both forms of offline simulation and online A/B testing. In the recent production launch, our system yielded 3-5 folds improvement over the existing production system in click-through, purchase-through and gross merchandising value; thus now driving 100% related recommendation traffic with billions of items at eBay. We believe that this work provides a practical yet principled framework for recommendation in the domains with affluent user self-input data.
- D. Agarwal and B.-C. Chen. Regression-based latent factor models. ACM Conference on Knowledge Discovery and Data Mining (KDD 2009), pages 19--28, 2009. Google ScholarDigital Library
- D. Agarwal and B.-C. Chen. fLDA: Matrix factorization through latent Dirichlet allocation. ACM International Conference on Web Search and Data Mining (WSDM 2010), 2010. Google ScholarDigital Library
- J. Bennett and S. Lanning. The Netflix Prize. In KDDCup'07: Proceedings of KDD Cup and Workshop 2007, 2007.Google Scholar
- D. M. Blei, A. Y. Ng, and M. I. Jordan. Latent Dirichlet allocation. Journal of Machine Learning Research, 3:993--1022, 2003. Google ScholarDigital Library
- O. Chapelle and Y. Zhang. A dynamic Bayesian network click model for web search ranking. International World Wide Web Conference (WWW 2009), pages 1--10, 2009. Google ScholarDigital Library
- Y. Chen and J. F. Canny. Probabilistic clustering of an item. U.S. Patent Application 12/694,885, filed: Jan, 2010.Google Scholar
- Y. Chen and J. F. Canny. Probabilistic recommendation of an item. U.S. Patent Application 12/694,903, filed: Jan, 2010.Google Scholar
- Y. Chen, M. Kapralov, D. Pavlov, and J. F. Canny. Factor modeling for advertisement targeting. Advances in Neural Information Processing Systems (NIPS 2009), 2009.Google Scholar
- Y. Chen, D. Pavlov, P. Berkhin, A. Seetharaman, and A. Meltzer. Practical lessons of data mining at yahoo! ACM Conference on Information and Knowledge Management (CIKM 2009), 2009. Google ScholarDigital Library
- Y. Chen, D. Pavlov, and J. F. Canny. Large-scale behavioral targeting. ACM Conference on Knowledge Discovery and Data Mining (KDD 2009), pages 209--218, 2009. Google ScholarDigital Library
- N. Craswell, O. Zoeter, M. Taylor, and B. Ramsey. An experimental comparison of click position-bias models. Web Search and Web Data Mining (WSDM 2008), pages 87--94, 2008. Google ScholarDigital Library
- R. O. Duda, P. E. Hart, and D. G. Stork. Pattern Classification. Wiley-Interscience, 2nd Edition (October 2000), 2000. Google ScholarDigital Library
- F. Guo, C. Liu, A. Kannan, T. Minka, M. Taylor, Y.-M. Wang, and C. Faloutsos. Click chain model in web search. International World Wide Web Conference (WWW 2009), pages 11--20, 2009. Google ScholarDigital Library
- T. Hofmann. Probabilistic latent semantic indexing. ACM Conference on Information Retrieval (SIGIR 1999), pages 50--57, 1999. Google ScholarDigital Library
- T. Hofmann and J. Puzicha. Latent class models for collaborative filtering. International Joint Conference on Artificial Intelligence (IJCAI 1999), pages 688--693, 1999. Google ScholarDigital Library
- F. Jelinek and R. L. Mercer. Interpolated estimation of Markov source parameters from sparse data. Pattern Recognition in Practice, pages 381--402, 1980.Google Scholar
- Y. Koren. Factorization meets the neighborhood: a multifaceted collaborative filtering model. ACM Conference on Knowledge Discovery and Data Mining (KDD 2008), pages 426--434, 2008. Google ScholarDigital Library
- Y. Koren. Collaborative filtering with temporal dynamics. ACM Conference on Knowledge Discovery and Data Mining (KDD 2009), pages 447--456, 2009. Google ScholarDigital Library
- A. Mccallum and K. Nigam. A comparison of event models for naïve Bayes text classification. AAAI-98 Workshop on Learning for Text Categorization, 1998.Google Scholar
- K. Nigam, A. K. McCallum, S. Thrun, and T. M. Mitchell. Text classification from labeled and unlabeled documents using EM. Machine Learning, 39(2):103--134, 2000. Google ScholarDigital Library
- D. Pavlov, R. Balasubramanyan, B. Dom, S. Kapur, and J. Parikh. Document preprocessing for naïve Bayes classification and clustering with mixture of multinomials. ACM Conference on Knowledge Discovery and Data Mining (KDD 2004), pages 829--834, 2004. Google ScholarDigital Library
- L. Si and R. Jin. Flexible mixture model for collaborative filtering. International Conference on Machine Learning (ICML 2003), pages 704--711, 2003.Google Scholar
- C. Zhai and J. Lafferty. A study of smoothing methods for language models applied to ad hoc information retrieval. ACM Conference on Information Retrieval (SIGIR 2001), pages 343--348, 2001. Google ScholarDigital Library
- S. Zhong and J. Ghosh. A comparative study of generative models for document clustering. SIAM International Conference Data Mining Workshop on Clustering High Dimensional Data and Its Applications, 2003.Google Scholar
Index Terms
- Recommending ephemeral items at web scale
Recommendations
CFSF: On Cloud-Based Recommendation for Large-Scale E-commerce
Recommender systems assist the e-commerce providers for services computing in aggregating user profiles and making suggestions tailored to user interests from large-scale data. This is mainly achieved by two primary schemes, i.e., memory-based ...
Learning from Sets of Items in Recommender Systems
Most of the existing recommender systems use the ratings provided by users on individual items. An additional source of preference information is to use the ratings that users provide on sets of items. The advantages of using preferences on sets are ...
A hybrid knowledge-based approach to collaborative filtering for improved recommendations
Collaborative filtering (CF) is one of the most successful and effective recommendation techniques for personalized information access. This method makes recommendations based on past transactions and feedback from users sharing similar interests. ...
Comments