摘要
研究了智能答疑系统中的问题分类。针对基于章节目录的分类方式过于依赖特定教材的不足,提出了基于关键词聚类的问题模糊分类方法。此方法基于关键词的语义,采用NERF算法对关键词进行聚类。并利用聚类有效性的方法来弥补此算法过于依赖初始值的不足。最后通过实例进行分析,说明此分类方法的可行性和对基于章节目录的分类方式不足的弥补。
The paper researches the question classification for intelligent question answering system. Aimed at the lack of the old classification algorithm, the paper presents a question fuzzy classification algorithm based on clustering of .Cluster keywords using non-Euclidean relational fuzzy c-means algorithm based the semantic of keywords, and use clustering validity to make up the shortage of relying on initializtion.Finally, analyse it through an example,and explain that the algorithm is feasible.
出处
《微机发展》
2005年第2期69-72,共4页
Microcomputer Development
关键词
智能答疑系统
问题模糊分类
聚类有效性
intelligent question answering system
question fuzzy classification
clustering validity
