ExpertFOAF recommends experts

FOAF files are often used for publishing simple information about persons and about their community. Our proposal is to extend the user’s FOAF files with automatically generated user profiles, which are based on a histogram of the user’s interest and which are also semantically enriched using the GrowBag approach [1]. The main assumption is that such user profiles can provide good hints about users’ expertise. Such extended FOAF files (called ExpertFOAF) can be published on a user’s home page, on web pages of institutions or conferences to characterize them. They can be crawled by distributed recommender systems for finding users with similar interests and, hence, expertise in different domains. We consider a well-defined user profile to express best the user’s current interests in a domain, where the domain (such as the research environment with students and professors) is defined by a collection of tagged objects (e.g., publications and deliverables tagged with keywords).