Sciweavers

Share
WWW
2010
ACM

Factorizing personalized Markov chains for next-basket recommendation

9 years 3 months ago
Factorizing personalized Markov chains for next-basket recommendation
Recommender systems are an important component of many websites. Two of the most popular approaches are based on matrix factorization (MF) and Markov chains (MC). MF methods learn the general taste of a user by factorizing the matrix over observed user-item preferences. On the other hand, MC methods model sequential behavior by learning a transition graph over items that is used to predict the next action based on the recent actions of a user. In this paper, we present a method bringing both approaches together. Our method is based on personalized transition graphs over underlying Markov chains. That means for each user an own transition matrix is learned – thus in total the method uses a transition cube. As the observations for estimating the transitions are usually very limited, our method factorizes the transition cube with a pairwise interaction model which is a special case of the Tucker Decomposition. We show that our factorized personalized MC (FPMC) model subsumes both a com...
Steffen Rendle, Christoph Freudenthaler, Lars Schm
Added 14 May 2010
Updated 14 May 2010
Type Conference
Year 2010
Where WWW
Authors Steffen Rendle, Christoph Freudenthaler, Lars Schmidt-Thieme
Comments (0)
books