Z. Gantner, L. Drumond, C.
Freudenthaler & L. Schmidt-Thieme; JMLR W&CP 18:231–247, 2012.
Personalized Ranking for Non-Uniformly Sampled Items
We develop an adapted version of the Bayesian Personalized Ranking (BPR
criterion (Rendle et al., 2009) that takes the non-uniform sampling of negative test items — as in
track 2 of the KDD Cup 2011 — into account. Furthermore, we present a modiﬁed version of the generic BPR
learning algorithm that maximizes the new criterion. We use it to train ranking
matrix factorization models as components of an ensemble. Additionally, we combine
the ranking predictions with rating prediction models to also take into account rating
With an ensemble of such combined models, we ranked 8th (out of more than 300 teams) in
track 2 of the KDD Cup 2011, without using the additional taxonomic information oﬀered by the competition organizers.
Page last modified on Tue May 29 10:23:38 2012.