Explaining reviews and ratings with PACO: Poisson Additive Co-Clustering

PDF

Understanding a user's motivations provides valuable information beyond the ability to recommend items. Quite often this can be accomplished by perusing both ratings and review texts, since it is the latter where the reasoning for specific preferences is explicitly expressed. Unfortunately matrix factorization approaches to recommendation result in large, complex models that are difficult to interpret and give recommendations that are hard to clearly explain to users. In contrast, in this paper, we attack this problem through succinct additive co-clustering. We devise a novel Bayesian technique for summing co-clusterings of Poisson distributions. With this novel technique we propose a new Bayesian model for joint collaborative filtering of ratings and text reviews through a sum of simple co-clusterings. The simple structure of our model yields easily interpretable recommendations. Even with a simple, succinct structure, our model outperforms competitors in terms of predicting ratings with reviews.
Submitted 6 Dec 2015 to Learning [cs.LG]
Published 8 Dec 2015
Subjects: cs.LG stat.ML
http://arxiv.org/abs/1512.01845
http://arxiv.org/pdf/1512.01845.pdf

0 comments