Paper Title
Finding High Quality Answers using Multi-Layer Model in Community Question Answering

Abstract
In recent years, Community Question Answering (CQA) websites have become biggest source of information where information seekers ask questions and domain experts give specific answers for those questions. Though users can easily get targeted information from CQA websites, there are many challenges associated with it for keeping CQA websites growing as a source of knowledge. Categorizing questions, identifying duplication questions, predicting best answers and finding domain experts are few of them. Popular CQA websites like Stack Overflow, Quora, Yahoo! Answers, etc. allow user to select any one answer as accepted answer for any question. While sometimes for non-factoid or subjective questions, rather than considering any one answer as accepted / best answer, multiple answers can be characterized as high-quality answers. Even if a particular answer is not exactly applicable for questioner, it could be helpful for other information seeker. Most of CQA websites have multi-layer communication channel where user not only post answer for question but also write his/her views (comments) on answer. Visitors express their views on a specific answer through comment or voting. Hence comments’ data for any answer can be considered as important dataset for determining quality of corresponding answer. Most existing work focuses on predicting best answer using various features’ analysis related to Question-Answer pairs. In this paper, we propose a novel approach for classifying answers as high / low quality answers based on multi-layer model i.e. Question-Answer analysis, Answer-Answer analysis and Answer-Comments analysis. Keywords - Community Question Answering (CQA), Answer-Comment Analysis, Answer’s classification