A Graphical Model Formulation of Collaborative Filtering Neighbourhood Methods with Fast Maximum Entropy Training
This work addresses collaborative filtering for recommendation systems, offering an incremental improvement in efficiency and performance over existing neighborhood methods.
The paper tackles the problem of collaborative filtering by introducing a new neighborhood method called item fields, which uses an undirected graphical model over an item graph to incorporate non-local information, achieving best results with drastically fewer edges than other approaches. It presents a fast approximate maximum entropy training method that is two orders of magnitude faster than maximum likelihood approaches on Movielens datasets.
Item neighbourhood methods for collaborative filtering learn a weighted graph over the set of items, where each item is connected to those it is most similar to. The prediction of a user's rating on an item is then given by that rating of neighbouring items, weighted by their similarity. This paper presents a new neighbourhood approach which we call item fields, whereby an undirected graphical model is formed over the item graph. The resulting prediction rule is a simple generalization of the classical approaches, which takes into account non-local information in the graph, allowing its best results to be obtained when using drastically fewer edges than other neighbourhood approaches. A fast approximate maximum entropy training method based on the Bethe approximation is presented, which uses a simple gradient ascent procedure. When using precomputed sufficient statistics on the Movielens datasets, our method is faster than maximum likelihood approaches by two orders of magnitude.