A Multiagent Referral System for Expertise Location

A referral system is an agent-based framework for assisting and simplifying person-to-person communication , such as nding experts for a speci-ed topic. The informal person-to-person social networks are used as \referral chain" requests for expertise, and software agents help automate the search of information and expertise. However, current systems rely exclusively on collections of email messages or newsgroup postings. As a result , previous approaches yield poor referrals for requests outside of the domain of the given collection. By contrast, we develop an approach that (1) combines techniques from information retrieval, multiagent learning and adaptive user modeling to reene the social network according to the user's needs, and (2) allows agents to unin-trusively exchange explicit proole information to add coverage regarding a user's expertise. An experimental simulation has been implemented and results of its use are presented.