JISE


  [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13]


Journal of Information Science and Engineering, Vol. 28 No. 1, pp. 51-65


Semantic Link Analysis for Finding Answer Experts

YAO LU1,2,3, XIAOJUN QUAN2, JINGSHENG LEI4, XINGLIANG NI1,2,3, WENYIN LIU2,3 AND YINLONG XU1,3
1School of Computer Science and Technology 
University of Science and Technology of China 
Hefei, 230026 P.R. China 
2Department of Computer Science 
City University of Hong Kong 
HKSAR, P.R. China 
3Joint Research Lab of Excellence 
CityU-USTC Advanced Research Institute 
Suzhou, 215123 P.R. China 
4School of Computer and Information Engineering 
Shanghai University of Electronic Power 
Shanghai, 200090 P.R. China

    Recommending unanswered questions to answer experts is an important mechanism in User-Interactive Question Answering (UIQA) services and is helpful to reduce asker's waiting time and obtain high-quality answers. In this paper, we address the task of identifying answer experts in UIQA services with semantic information extracted from user interaction behaviors. We first construct the user question-answer interaction graph through direct semantic links and latent links extracted from the records of question sessions and user profiles. After that, two expert-finding approaches are developed by employing the semantic information in the so-called propagation link analysis method and in the language model, respectively. Experimental results on Yahoo! Answers dataset show that the extracted semantic information indeed improves the performance of both propagation and language model for the task of answer experts finding.

Keywords: user-interactive question answering, answer expert finding, semantics, link analysis, language model

  Retrieve PDF document (JISE_201201_04.pdf)