The "Spree" Expert Finding System

This paper proposes a system to facilitate exchange of information by automatically finding experts, competent in answering a given question. Our objective is to provide an online tool, which enables individuals within a potentially large organization to search for experts in a certain area, which may not be represented in company organization or reporting lines. The advantage of the proposed system over standard forums or group-ware systems is that full-formatted questions can be compared to stored qualification profiles, which were automatically derived from documents, without Human search effort, and possibly refined manually. This allows us to find competent colleagues (or helpful literature such as How-Tos) for a given problem in a single step, and without intermediate iterations. The system is symmetric in that it does not distinguish between "questioners" (asking questions) and "experts" (answering them), therefore forming a "community" of users, which are distributed over an ontology covering the total knowledge. This ontology can either be given (i.e. in the form of an organizational chart), or it can be derived from the experts\´ knowledge.