支持向量机在文本特征提取中的应用研究

ON APPLYING SUPPORT VECTOR MACHINE IN TEXT FEATURE EXTRACTION

在文本自动分类中,针对如何进行文本特征的选择和提取这一关键和基础性工作,提出用支持向量度量词汇对分类的贡献,然后进行文本特征的提取。实验结果表明,该方法可以在确保分类信息不损失的前提下,降低向量空间的维数,提高分类器效率和分类准确率。

Aiming at the key and foundation work of selecting and extracting text feature in automatic text classification,a method of extracting text feature with Support Vector Machine was proposed.It uses the SVM to measure the contribution of the vocabulary on classification and then to extract the text feature.Experimental results show that this method can reduce the dimension of the vectors space and improve the efficiency of classifier and the accuracy of classification while ensuring the information of classification lossless.