Community-based Question Answering (CQA) sites, such as Yahoo! Answers, Baidu Knows, Aardvark and Quora, provide an alternative way of online information seeking other than web search, especially for hard, complex, and personal information needs that are not well serviced by web search engines. Additionally, a large number of web searchers could also benefit from CQA sites by finding existing answers that address their information needs via search engines. Moreover, the interaction of web searchers with CQA sites could provide useful hints for better understanding what they are actually searching for, e.g., either the questions they clicked on in
successful searches or the questions they posted on CQA sites after unsuccessful searches contain better descriptions of what they are actually searching for, than the search queries alone.
successful searches or the questions they posted on CQA sites after unsuccessful searches contain better descriptions of what they are actually searching for, than the search queries alone.
Based on these observations, this project proposes to use CQA services and related data to better help users address difficult information needs. We identify three challenging research questions towards this goal:
- What are the factors influencing answer contributions in CQA systems?
- What kind of searches benefit most from CQA services and archives?
- How to improve search quality with CQA data for difficult information needs?
By answering the proposed research questions, this project improves the quality of CQA services and offers new ways of using CQA data to improve search.