摘要
在介绍文本分类的背景及传统基于向量空间模型特征选择不足之处的同时,提出了不同特征选择方法相结合的文本分类模型.该模型首先对文本进行分析,把文本表示成向量空间的形式.文本在经过预处理后,按一定规则提取关键词,关键词的提取中增加了对名词短语的识别.特征选择的方法上,结合了文档频数和互信息量,并对他们进行了改进.实验结果表明,使用新方法进行分类所得到的分类精度得到了一定的提高.
A new text categorization system using composite feature selection method was designed.Texts are described in the form of Vector Space Model through model analysis.After pretreatment for texts, which include noun phrase were extracted on the basis of rules from texts.On feature selection,document frequency was combined with mutual information,and performance was improved.Experiments show that the precision of text categorization is improved through using new method.
出处
《厦门大学学报(自然科学版)》
CAS
CSCD
北大核心
2005年第B06期239-242,共4页
Journal of Xiamen University:Natural Science