摘要
提出了一种基于DOM树的词共现模型,首先利用文档的结构信息生成DOM树,并依据DOM树的结构特点来统计文档中主题词的共现信息,最后采用向量空间模型实现对石油主题网页的采集和分类。它改进了原有的词共现模型,突出了利用位置信息来优化词共现模型的特点。实验证明该策略使采集和分类的性能都有了一定的提高。
This paper proposes a word co - occurrence model which bases on the DOM tree. At first , the model uses the structure information of the document to build a DOM tree, and then counts the co - occurrence information in the documents bases on the structure of the DOM tree ,finally implements collection and categorization for the petroleum subject. It improves the original word co - occur- rence model and highlighted the character of positional information to optimize the word co - occurrence model. The experiment proves this strategy caused a certain enhancement for collection and categorization.
出处
《微计算机应用》
2008年第2期28-31,共4页
Microcomputer Applications
关键词
词共现模型
DOM树
文本分类
主题采集
向量空间模型
word co-occurrence model, DOM(document object model)tree, text categorization, focused web crawling, vector space model