Article Preview
Top1. Introduction
Question answering is a computer science discipline within the fields of information retrieval and natural language processing. It is the task of automatically answering a question posed in natural language. QA systems deliver users short, succinct answers instead of overloading them with a large number of irrelevant documents (Lin and7 Katz 2003). This is the goal of every QA system. More commonly, QA systems pull answers from unstructured collection of natural language document. Community Question Answering (CQA) services are dedicated platforms for users to respond to other users’ questions, resulting in the building of social communities where users share and interactively give ratings to questions and answers (Wang and7 Zhang 2011; Pal and7 Konstan 2010). From observation, more users are becoming inclined to this new area of QA systems as they can obtain a near perfect answer to their questions from other users rather than a list of likely documents containing result(s) to their questions that are provided by the system. While current web search engines enjoy huge commercial success and demonstrate good performance, especially for homepage-finding queries, their ability to find relevant information for hard queries such as those asking for opinions or summaries is far from satisfactory (Jeon et al., 2005).
Many of the existing works on QA systems have focused on retrieving high quality answers to an asker’s question and recommending answerers (Hieber and7 Riezler 2011). However, the time lag that results from askers having to wait for answers to a question from various users as identified by Jeon et al. (2005) have not received much consideration. In addition, while browsing the internet, we often spend large amount of time searching for a particular answer that best answers our question, and end up getting too large materials that are not specific and so difficult to extract specific and quality answers to the questions we have. If we later extract some line of text or a paragraph of text we are not even sure if the extracted text is the best, in this case an interactive question answering system that enables the asker get specific community rated expert answers would have been better. Answering communities such as Yahoo! Answers offer great intelligence to the users who have questions in either their daily lives or academic affairs. Discovering experts in these communities is a very important research problem that have been explored by a number of researchers (Liu et al., 2005; Jurczyk and7 Agichtein, 2007; Pal et al., 2012). Participants can express their judgments towards answers by voting for the answer they feel is the best among the expert provided answers. The need to obtain aggregate opinions from these experts relating to a particular question in an effective way forms one of the motivations for this research work.