Expert Finding in CQA Based on Topic Professional Level Model

• Published in: Data Mining and Big Data Vol. 10943; pp. 459 - 465
• Main Authors: Wang, Shuaiyang, Jiang, Di, Su, Lei, Fan, Zhengyu, Liu, Xi
• Format: Book Chapter
• Language: English
• Published: Switzerland Springer International Publishing AG 2018; Springer International Publishing
• Series: Lecture Notes in Computer Science
• Subjects: Community Question Answering; Expert finding; Topic Professional Level Model
• ISBN: 3319938029; 9783319938028
• ISSN: 0302-9743; 1611-3349
• DOI: 10.1007/978-3-319-93803-5_43

In the CQA (Community Question Answering) systems, expert finding is one of the most important subjects. The task of expert finding is aimed at discovering users with relevant expertise or experience for a given question. However, with the increasing amount of information in CQA platform, the questioner has to wait for a long time for the response of other users, and the quality of the answers that user receive is not optimistic. In view of the above problems, this paper proposes the Topic Professional Level Model (TPLM) to find the right experts for questions. The model combines both the topic model and the professional level model respectively from the two perspectives of semantic topic of textual content and link structure to calculate the user’s authority under a specific topic. Based on TPLM results, this paper proposed the TPLMRank algorithm to measure user comprehensive score to find the expert users. The experimental results on the Chinese CQA platform-Zhihu dataset show that the expert finding method based on the TPLM is superior to the traditional expert finding method.