社区问答系统已经成为获取和分享知识的一种重要渠道,但用户提供的信息质量差异比较大。针对社区问答系统中具有多个答案的问题,提出了一种基于混合式的社区问答答案质量评价模型,可实现最佳答案的选取。该模型首先利用基于用户活动的UAM模型获得问题和答案的主题相似度并剔除无关回复,然后结合用户权威度及多重评价标准,对答案进行评分,获得对答案的定量评价结果。基于Stack Overflow的实验表明,该方法可有效地对答案进行质量评分,有实用价值。
Community question answering had become an important channel of getting and sharing knowledge, but the user information quality is unevenness. The paper specific to the question which had multiple answers, proposed an answer quality evaluation model based on hybrid and chosen the best answer. Firstly, the model used UAM which was based on user' s activity to get the topic similarity of question and answer, eliminated irrelevant answers ; then combined with user authority and multiple evaluation standards, gave the left answers a score and get a quantitative result. Based on the test Of Stack Overflow, this model can grade the answer quality effectively and it is practical.