Learning to Rank Experts in Academic Digital Libraries

The task of expert finding has been getting increasing attention in the information retrieval literature. However, the current state-of-the-art still lacks in principled approaches for combining different sources of evidence in an optimal way. This article explores the usage of learning to rank methods as a principled approach for combining multiple estimators of expertise, derived from the textual contents, from the graph-structure of the citation patterns for the community of experts, and from profile information about the experts. Several supervised learning to rank algorithms, representative of the pointwise, pairwise and the listwise approaches, were experimented. This article also explores the effectiveness of rank aggregation approaches combined with data fusion techniques in the task of expert finding in digital libraries. Several experiments were performed with state of the art data fusion techniques. Experiments made over a dataset of academic publications for the Computer Science domain attest for the adequacy of the proposed approaches.