Authors : Jae-Wook Seol, Seok-Hyoung Lee, Seo-Young Jeong, Hye-Jin Lee, Jeong-Seon Yoon and Kwang-Young Kim
Abstract: In the modern society, unexpected events such as diseases and disasters are advertent. In specific specialized sector, finding of expert is important for resolving social issues. This study proposes that expert finding model for each sector through quantification of expertise of researchers. First, the issue of ambiguity of author in academic data shall be resolved. To measure precisely the expertise of an author we conduct author identification in which the author name written in different forms is identified as an actual personnel. Second, based on the accumulated importance of author keyword and reference network, we apply the modified Hyperlink-Induced Topic Search (HITS) algorithm to extract out expert candidates. To verify the validity of this proposal, expert finding shall be conducted on 92,100 cases of academic data incurred in Korea. We evaluate our expert finding model based on human relevance judgments about several queries. The outcome of experimenting author importance resolution is F1 measure 94.79% and the expert finding model applied with our modified HITS algorithm shows the mean average precision of 75%.
Jae-Wook Seol, Seok-Hyoung Lee, Seo-Young Jeong, Hye-Jin Lee, Jeong-Seon Yoon and Kwang-Young Kim, 2017. Expert Finding Model Through Author Disambiguation in Bibliographic Data. Journal of Engineering and Applied Sciences, 12: 2597-2602.