Loading...

Expert Finding in Community Question Answering

Miri, Mohammad | 2021

249 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 54482 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Beigy, Hamid
  7. Abstract:
  8. Community question answering are the systems in which users can propose their needs with asking questions. Moreover, they can share their own knowledge with the others by responding their questions. Widely spreading this sort of communities and growth of questions and answers has made some challenges. One of these challenges is finding an appropriate users who can answer questions. For instance, some user might ask a question and has to wait some times to receive another user's response. On the other hand, the users who have expertise in some fields have to spend time a lot seeking to find related questions. Therefore, expert finding systems are used to meet these needs.The main issue is semantic gap between search queries and text of questions and answers in the system. In this research, we propose two translation models which are using word embedding. In these approaches queries and words are represented as vectors of latent variables. First, we get similarity between the existing words in CQA and the query, then search the most appropriate users to respond the query. Finally, the users with most use of translated words are the best candidate to respond the query.To measure the proposed models accuracy we compare them to the best existing methods. The dataset is used in this research is Stackoverflow dataset. Furthermore, we use MAP measure to compare them.
  9. Keywords:
  10. Expert Finding ; Community Question Answering ; Improving Answer in Stack Overflow

 Digital Object List

 Bookmark

No TOC