We present a clustering method designed to complement existing model-based collaborative filtering algorithms. Thanks to clustering, we are able to produce faster recommendations on datasets with a very large number of items. Moreover, we can work with models independent of the number of users, and produce recommendations for new users and in session based environment. We show that our method builds meaningful clusters, and that we can reduce the number of items considered during recommendation by a factor up to 50, with little impact on the accuracy or diversity of the recommendation.

Recurrent neural networks have recently been successfully applied to the session-based recommendation problem, and is part of a growing interest for collaborative filtering based on sequence prediction. This new approach to recommendations reveals an aspect that was previously overlooked: the difference between short-term and long-term recommendations. In this work we characterize the full short-term/long-term profile of many collaborative filtering methods, and we show how recurrent neural networks can be steered towards better short or long-term predictions. We also show that RNNs are not only adapted to session-based collaborative filtering, but are perfectly suited for collaborative filtering on dense datasets where it outperforms traditional item recommendation algorithms.

Advanced and effective collaborative filtering methods based on explicit feedback assume that unknown ratings do not follow the same model as the observed ones not missing at random). In this work, we build on this assumption, and introduce a novel dynamic matrix factorization framework that allows to set an explicit prior on unknown values. When new ratings, users, or items enter the system, we can update the factorization in time independent of the size of data (number of users, items and ratings). Hence, we can quickly recommend items even to very recent users. We test our methods on three large datasets, including two very sparse ones, in static and dynamic conditions. In each case, we outrank state-of-the-art matrix factorization methods that do not use a prior on unknown ratings.

Although criticized for some of its limitations, modularity remains a standard measure for analyzing social networks. Quantifying the statistical surprise in the arrangement of the edges of the network has led to simple and powerful algorithms. However, relying solely on the distribution of edges instead of more complex structures such as paths limits the extent of modularity. Indeed, recent studies have shown restrictions of optimizing modularity, for instance its resolution limit. We introduce here a novel, formal and well-defined modularity measure based on random walks. We show how this modularity can be computed from paths induced by the graph instead of the traditionally used edges. We argue that by computing modularity on paths instead of edges, more informative features can be extracted from the network. We verify this hypothesis on a semi-supervised classification procedure of the nodes in the network, where we show that, under the same settings, the features of the random walk modularity help to classify better than the features of the usual modularity. Additionally, the proposed approach outperforms the classical label propagation procedure on two data sets of labeled social networks.