Arnetminer: Expertise Oriented Search Using Social Networks
Juanzi Li,Jie Tang,Jing Zhang,Qiong Luo,Yunhao Liu,Mingcai Hong
DOI: https://doi.org/10.1007/s11704-008-0008-9
2008-01-01
Frontiers of Computer Science in China
Abstract:Expertise Oriented Search (EOS) aims at providing comprehensive expertise analysis on data from distributed sources. It is useful in many application domains, for example, finding experts on a given topic, detecting the confliction of interest between researchers, and assigning reviewers to proposals. In this paper, we present the design and implementation of our expertise oriented search system, Arnetminer (http://www.arnetminer.net). Arnetminer has gathered and integrated information about a half-million computer science researchers from the Web, including their profiles and publications. Moreover, Arnetminer constructs a social network among these researchers through their co-authorship, and utilizes this network information as well as the individual profiles to facilitate expertise oriented search tasks. In particular, the co-authorship information is used both in ranking the expertise of individual researchers for a given topic and in searching for associations between researchers. We have conducted initial experiments on Arnetminer. Our results demonstrate that the proposed relevancy propagation expert finding method outperforms the method that only uses person local information, and the proposed two-stage association search on a large-scale social network is order of magnitude faster than the baseline method.