Learning multiple models for exploiting predictive heterogeneity in recommender systems

Abstract

Collaborative filtering approaches exploit information about historical affinities or ratings to predict unknown affinities between sets of "users" and "items" and make recommendations. However a model that also incorporates heterogeneous sources of information that may be available on the users and/or items can become a much more effective recommender, in terms of both increased relevance of the predictions as well as explainability of the results. In this paper, we propose a Bayesian approach that exploits not only such "side-information", but also a different kind of heterogeneity that captures the variations in the mapping from user/item attributes to the affinities of interest. Such predictive heterogeneity is likely to occur in large recommender systems that involve a diverse set of users, and can be mitigated by using multiple localized predictive models rather than a single global one that covers all user-item pairs. The scope or coverage of each local model is determined simultaneously with the model parameters. The proposed approach can incorporate different types of inputs to predict the preferences of diverse users and items. We compare it against well-known alternative approaches and analyze the results in terms of both accuracy and interpretability.

Using Tags for Measuring the Semantic Similarity of Users to Enhance Collaborative Filtering Recommender Systems

“…It is based on the original MovieLens10M dataset, published by the Group Lens1 research group. This dataset has been used in previous studies such as [32], [33]. One of the main difficulties when coping with tagging data is the quality of the tags mainly because tags are words or a combination of words that are that are freely assigned by users.…”

Integrating heterogeneous information via flexible regularization framework for recommendation

“…More and more researchers have been aware of the importance of heterogeneous information for recommendation. Jones et al [8] validated the importance of the exploitation on available heterogeneous data sources and proposed a Bayesian approach called LaD-BAE to capture both feature heterogeneity and predictive heterogeneity. Zhang et al [28] investigated the problem of recommendation over heterogeneous network and formalized the recommendation as a ranking problem then proposed a random walk model to estimate the importance of each object in the heterogeneous network.…”