作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2011, Vol. 37 ›› Issue (14): 70-72. doi: 10.3969/j.issn.1000-3428.2011.14.022

• 软件技术与数据库 • 上一篇    下一篇

基于概率潜在语义分析模型的自动答案选择

张 成 1,曲明成 2,倪 宁 3,仇 光 2,卜佳俊 2   

  1. (1. 中国残联信息中心,北京 100034;2. 浙江大学计算机科学与技术学院,杭州 310027;3. 浙江商业职业技术学院信息技术学院,杭州 310053)
  • 收稿日期:2011-02-14 出版日期:2011-07-20 发布日期:2011-07-20
  • 作者简介:张 成(1970-),男,硕士,主研方向:数据挖掘,语义分析;曲明成,硕士;倪 宁,副教授;仇 光,博士;卜佳俊,教授
  • 基金资助:
    国家科技支撑计划基金资助项目(2008BAH26B00)

Automatic Answer Selection Based on Probabilistic Latent Semantic Analysis Model

ZHANG Cheng 1, QU Ming-cheng 2, NI Ning 3, QIU Guang 2, BU Jia-jun 2   

  1. (1. China Disabled Persons’ Federation Information Center, Beijing 100034, China; 2. College of Computer Science and Technology, Zhejiang University, Hangzhou 310027, China;3. College of Information Technology, Zhejiang Vocational College of Commerce, Hangzhou 310053, China)
  • Received:2011-02-14 Online:2011-07-20 Published:2011-07-20

摘要: 问答社区中候选答案过多会增加提问用户选择最佳答案的负担。为此,提出一种基于概率潜在语义分析(PLSA)模型的自动答案选择方法。在主题建模思想的基础上,利用问答社区中的用户资料,以PLSA模型表达问答社区中的用户兴趣分布,依据答案和问题之间的主题匹配度对候选答案进行排序。实验结果表明,该方法可有效挖掘用户兴趣,提高答案选择的准确率。

关键词: 答案选择, 问答社区, 概率潜在语义分析, 主题建模

Abstract: A novel answer selection method based on topic modeling techniques is proposed to mitigate the issue of question asker’s burden of selecting the best answer stemming from too many candidate answers in question answering communities. Aiming at the problem, this paper presents an automatic answer selection based on Probabilistic Latent Semantic Analysis(PLSA) in question answering communities, and accordingly rank candidate answers based on similarity of interest between answers and questions. Experimental results show that the method can effectively excavation user interest and improve the accuracy of answer selection.

Key words: answer selection, question & answering community, Probabilistic Latent Semantic Analysis(PLSA), topic modeling

中图分类号: