论文信息 - Improving Incremental Recommenders with Online Bagging

Improving Incremental Recommenders with Online Bagging

Online recommender systems often deal with continuous, potentially fast and unbounded flows of data. Ensemble methods for recommender systems have been used in the past in batch algorithms, however they have never been studied with incremental algorithms that learn from data streams. We evaluate online bagging with an incremental matrix factorization algorithm for top-N recommendation with positive-only user feedback, often known as binary ratings. Our results show that online bagging is able to improve accuracy up to 35% over the baseline, with small computational overhead.

João Gama | Alípio Mário Jorge | João Vinagre

[1] BoostMF: Boosted Matrix Factorisation for Collaborative Ranking , 2015, ECML/PKDD.

[2] Leo Breiman,et al. Bagging Predictors , 1996, Machine Learning.

[3] João Gama,et al. Fast Incremental Matrix Factorization for Recommendation with Positive-Only Feedback , 2014, UMAP.

[4] Joseph Sill,et al. Feature-Weighted Linear Stacking , 2009, ArXiv.

[5] Amnon Meisels,et al. Ensemble methods for improving the performance of neighborhood-based collaborative filtering , 2009, RecSys '09.

[6] Yoav Freund,et al. Experiments with a New Boosting Algorithm , 1996, ICML.

[7] João Gama,et al. On evaluating stream learning algorithms , 2012, Machine Learning.

[8] Robert A. Legenstein,et al. Combining predictions for accurate recommender systems , 2010, KDD.

[9] Geoff Hulten,et al. Catching up with the Data: Research Issues in Mining Data Streams , 2001, DMKD.

[10] David H. Wolpert,et al. Stacked generalization , 1992, Neural Networks.

[11] Lars Schmidt-Thieme,et al. MyMediaLite: a free recommender system library , 2011, RecSys '11.

[12] María N. Moreno,et al. An experimental comparative study of web mining methods for recommender systems , 2006 .

[13] Stuart J. Russell,et al. Experimental comparisons of online and batch versions of bagging and boosting , 2001, KDD '01.