基于向量空间模型的多主题Web文本分类方法*

对给定的网页,提取其特征向量,计算网页特征向量与分类特征向量的相似度,使用K-means聚类方法寻找归属类得到动态阈值,提出了一种基于动态阈值的向量空间模型多主题Web文本分类方法.该方法通过网页与每个类的相似度和动态阈值的比较,实现了将包含多个主题的网页划分到相应的多个类中.实验证明,这种方法具有较好的精确度和召回率.

Method of multi topic Web text classification based on VSM

Withdrawing characteristic vectors for a given Web page,calculating the similarities of the page characteristic vectors with classification characteristic vectors, getting dynamic thresholds through using K-means clustered methods and looking for result classifications, this paper proposed a multi-topic Web text classification method of vector space model based on dynamic threshold. Through comparing the value of every classification similarity with dynamic threshold, classifyed the multitopic texts of a Web page to several different text classifications. The simulating experiments verify the good accuracy and better recalling with this method.