一种基于文本分类的知识树自动构建方法  

Automatic construction of knowledge tree based on text clustering

作　　者：钟将[1] 刘杰[1] 

机构地区：[1]重庆大学计算机学院,重庆400044

出　　处：《计算机应用研究》2010年第2期475-478,共4页Application Research of Computers

基　　金：国家自然科学基金资助项目(60803027);重庆市科委自然科学基金资助项目(2008BB2195)

摘　　要：针对当前知识管理系统中知识树的创建和维护问题,设计了一种新的基于文本聚类的知识树构建方法。由于从传统的K-means和SOM等文本聚类的结果中难以提取知识树中节点对应的概念和词汇列表,选取PLSA方法进行聚类和知识层次树构建。实验表明,新方法除了在聚类精确度上优于传统方法,聚类结果还包含文档的主题与词汇之间的概率关系,因此新方法在聚类的同时,可以方便地提取知识树上每个节点对应的概念或概念集合。The construction and maintenance of the knowledge tree is an important and time-consuming task in a knowledge management system （KMS）. This paper presented a novel method to construct the knowledge tree based on text clustering. Because it＇ s difficult to extract concepts and vocabulary corresponding to nodes in knowledge tree while clustering by traditional K-means and SOM algorithms, selected PLSA （probabilistic latent semantic analysis） to construct knowledge tree. Experiment shows that the clustering accuracy of the new method is higher than the traditional K-means and SOM algorithms. In addition, because the probabilistic relationship between the vocabulary and the concept （subject） has been established, the concepts of node in knowledge tree could be easily extracted while clustering documents by the new method.

关 键 词：概率潜在语义分析 潜在语义空间 知识管理 知识树 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]