A Bayesian Approach to User Profiling in Information Retrieval

Numerous probability models have been suggested for information retrieval (IR) over the years. These models have been applied to try to manage the inherent uncertainty in IR, for instance, document and query representation, relevance feedback, and evaluating the effectiveness of IR system. On the other hand, Bayesian networks have become an established probabilistic framework for uncertainty management in artificial intelligence. In this paper, we suggest the use of Bayesian networks for user profiling in IR. Our approach can take full advantage of both the effective learning algorithms and efficient query processing techniques already developed for probabilistic networks. Moreover, Bayesian networks capture a more general class of probability distributions than the previously proposed probabilistic models. Finally, this paper provides a theoretical foundation for the cross-fertilization of techniques between IR and Bayesian networks.