A Service Mode of Expert Finding in Social Network
Xiu Li, Jianguo Ma, Yujiu Yang, Dongzhi Wang · 2013
Expert finding addresses the task of finding the right person with the appropriate knowledge or skills. State-of-the-art expert finding algorithms usually estimate the relevance between the query and the support documents of candidates using language model. However, the language model has a limitation that all the query terms should occur in each support document, which results in some real experts cannot be searched. So for the process of analyzing textual content, we consider using a new model based on Explicit Semantic Analysis (ESA) rather than the language model. With the development of Internet technology, this is not the only way to find experts. In the modern social media, we can record person's social relationships which might be available for expert finding task. A simple truth is: a person's connections with experts will provide the potential evidence that he is a real expert. In this paper, we propose a new service pattern for expert finding that accounts for both documents' content and social relationships. The relationships in the social network are used in re-ranking experts on a given topic.