Using transactional data from ERP systems for expert finding

Abstract

During the past decade, information retrieval techniques have been augmented in order to search for experts and not just documents. This is done by searching document collections for both query topics and associated experts. A typical approach assumes that expert candidates are authors of intranet documents, or that they engage in social writing activities on blogs or online forums. However, in many organizations, the actual experts, i.e., the people who work on problems in their day-to-day work, rarely engage in such writing activities. As an alternative, we turn to structured corporate data-transactions of working hours provided by an organization's ERP system-as a source of evidence for ranking experts. We design an expert finding system for such an enterprise and conclude that it is possible to utilize such transactional data, which is a result of required daily business processes, to provide a solid source of evidence for expert finding. © 2010 Springer-Verlag.