Non-local evidence for expert finding

The task addressed in this paper, finding experts in an enterprise setting, has gained in importance and inter-est over the past few years. Commonly, this task is ap-proached as an association finding exercise between people and topics. Existing techniques use either documents (as a whole) or proximity-based techniques to represent candi-date experts. Proximity-based techniques have shown clear precision-enhancing benefits. We complement both docu-ment and proximity-based approaches to expert finding by importing global evidence of expertise, i.e., evidence ob-tained using information that is not available in the imme-diate proximity of a candidate expert’s name occurrence or even on the same page on which the name occurs. Exam-ples include candidate priors, query models, as well as other documents a candidate expert is associated with. Using the CERC data set created for the TREC 2007 En-terprise track we identify examples of non-local evidence of expertise. We then propose modified expert retrieval mod-els that are capable of incorporating both local (either doc-ument or snippet-based) evidence and non-local evidence of expertise. Results show that our refined models significantly outperform existing state-of-the-art approaches