Computer Engineering and Applications ›› 2008, Vol. 44 ›› Issue (21): 133-137.DOI: 10.3778/j.issn.1002-8331.2008.21.037

• 机器学习 • Previous Articles     Next Articles

Automatic identifying query interfaces of deep Web with maximum entropy classifier

FANG Wei1,2,HUANG Li1,2,CUI Zhi-ming1,2   

  1. 1.Jiangsu Key Laboratory of Computer Information Processing Technology,Soochow University,Suzhou,Jiangsu 215006,China
    2.Institute of Intelligent Information Processing and Application,Soochow University,Suzhou,Jiangsu 215006,China
  • Received:2008-04-30 Revised:2008-06-06 Online:2008-07-21 Published:2008-07-21
  • Contact: FANG Wei

基于最大熵分类器的Deep Web查询接口自动判定

方 巍1,2,黄 黎1,2,崔志明1,2   

  1. 1.江苏省计算机信息处理技术重点实验室,江苏 苏州 215006
    2.苏州大学 智能信息处理及应用研究所,江苏 苏州 215006
  • 通讯作者: 方 巍

Abstract: Tremendous high-quality web information is deeply hidden in the Web,which can not be indexed by traditional search engines,so we call them Deep Web.Since query interface is the only entrance to the Deep Web,we must distinguish query interfaces of Deep Web.Since the Maximum Entropy Model could integrate various correlative and irrelative probability knowledge,it could deal with many problem well.So we use Maximum Entropy Model for query interface categorization in this paper.Compared with Bayes,C4.5 and SVM,Maximum Entropy shows its high quality.Moreover,it is useful to query interface categorization.

Key words: deep web, Html form, feature extraction, maximum entropy model

摘要: Web中包含着海量的高质量信息,它们通常处在网络深处,无法被传统搜索引擎索引,将这样的资源称为Deep Web。因为查询接口是Deep Web的唯一入口,所以要获取Deep Web信息就必须判定哪些网页表单是Deep Web查询接口。由于最大熵模型可以综合观察到的各种相关或不相关的概率知识,对许多问题的处理都可以达到较好的结果。因此,基于最大熵模型的分类性能,利用最大熵分类算法自动判定查询接口。并通过实验,将最大熵分类法与其它常用分类方法进行了比较,结果显示它的分类性能优于Bayes方法和C4.5方法,与SVM方法相当,表明这是一种非常实用的查询接口分类方法。

关键词: Deep Web, 网页表单, 特征提取, 最大熵模型