Question and Answer Recommendation in Question Answering Communities
|Course||Applied Computer Technology|
|Keywords||Q \u0026 Community Problems Recommended Answer recommend Topic Modeling Link Analysis|
Today, users have become online interactive quiz community access to information and knowledge sharing important medium. Such as Yahoo! Answers, Baidu and other Q \u0026 A community site that has published tens of thousands of daily problems. However, with the growing amount of data Q community, users need to spend more time to find their own interest. Thus, the question asker need to wait longer to get the answer to that question. Meanwhile, the problem of the rapid growth in the number of candidate answers, answers uneven levels of quality, but also increased the questions the user to select the best answer burden. This Q \u0026 A community for recommendations and answer questions recommendation mechanism conducted in-depth research, designed to help users to ask questions and answer user access to information, thus enhancing Answers community knowledge sharing behavior. Problems to be solved recommended Recommend to users interested in the issue, so that the question can be answered as soon as possible. This paper argues that the user according to their topics of interest to select the appropriate question to answer, so the question of user interest topics and answers the need for a higher degree of correlation. Based on this, we propose a topic-based modeling method recommended ideological problems, take advantage of the rich user community Q \u0026 personalized information to probabilistic latent semantic analysis model to express the user interest in the community Q \u0026 distribution, and this calculation Problem recommendation list. Answer questions recommended for candidate answers for automatic sorting, so that users can more easily ask questions choose the best answer. This paper argues that the user will answer questions and questions of quality and relevance Choose the best answer. This paper presents an approach based on questions and answers between the degree of similarity and authoritative answers to users recommended method. In this method, the user quiz relationship building user link graph, in order to use the PageRank algorithm to estimate the degree of user authority. To calculate the similarity, considering the content of the questions and answers as well as the similarity of the user and to answer user questions similarity. Experimental results show that the proposed topic-based modeling problems recommended method can effectively mining user interest and thus recommended to be solved. The answer is recommended experimental results prove that considering the questions and answers as well as the similarity of content users and answer user questions the validity of similarity, and degree of authority by the user to measure the quality of the feasibility of the answer.