Expertise Network Discovery via Topic and Link Analysis in Online Communities

Abstract

Online communities have become important places for people to seek and share expertise. Yet with the increasing number of members and produced artifacts within the communities, it is challenging to find the influential experts who post topic-specific high-quality content. This paper presents an approach to discover expertise network in online communities based on textual information and social links. In addition to computing documents' topic-focus degree, the approach measures the quality of documents according to users' feedback behaviors and topic-specific influence of users who give feedback. In this way, user's expertise rank and social links are both considered to constitute expertise network. Experiments on real dataset have shown that our approach is effective to discover the meaningful expertise networks.