摘要： 提出了一个基于本体的异构文本分类系统,使用结构本体很好地消除了文本文档的结构上的差异,并将领域本体引入到分类系统中,使得分类更加准确、高效,分类的规则更易理解。在使用COSA算法提取有效概念的同时,也大大地减少了关键术语的数量,节省了运算开销。

关键词: 本体, 知识工程, 机器学习, 文本分类, 规则

Abstract: This paper brings out an ontology-based heterogeneous text categorization system, which exploits structure ontology to avoid the structure difference of text documents, and introduces domain ontology to this text categorization system. All the measures mentioned above make the system more accurate and more effective, the rules for categorization more understandable. When COSA algorithm is used to get meaningful concepts, the number of

中图分类号: 

• TP391.1