摘要
文章提出了一种基于本体论的文本特征提取方法。通过构建文本结构树来充分利用文本结构分析得到的信息 ,利用本体对领域知识的描述信息来分析特征词之间的关系 ,而且在特征权值的计算中提出了特征词统领长度的概念和计算方法。实验数据表明该方法提高了文本特征提取的准确性。
The paper proposed a new method of the feature extraction of Chinese text based on Ontology. The method can make full use of the structure information by constructing the text structure tree. The description about on domain knowledge make it available to analyze the relation of the key words by the Ontology. The paper also present the weight formula, and put forward a new concept named presidential length and its formula. The experimental results display the improvement of the veracity of feature extraction.
出处
《电脑与信息技术》
2005年第1期36-38,62,共4页
Computer and Information Technology
