基于页面标签的Web结构化数据抽取  

Structured Data Extraction Based on Web Page Tags

作　　者：任仲晟[1] 薛永生[1] 

机构地区：[1]厦门大学计算机科学系,厦门361005

出　　处：《计算机科学》2007年第10期133-136,共4页Computer Science

基　　金：国家自然科学基金(50474033);福建省自然科学基金(A0310008);福建省重点科技项目(2003H043)

摘　　要：本文研究了从data intensive类型的Web页面中提取结构化数据的问题,提出了基于页面标签的数据抽取算法。该算法先根据标签的显示位置及其大小判断不同标签元素之间的嵌套关系,并构造简化的HTML树Sim- HTree,有效地减少了识别数据记录的时间。在此基础上,提出子串匹配调整算法,对数据记录进行识别,标识教据项。实验表明,该算法是有效的。This paper studies the problem of structured data extraction from data intensive Web pages. A novel approach based on page tags is proposed to solve the problem. The proposed method identifies the nesting relationship between different page tags according to the visual display location and the size on the screen and constructs the corresponding simplified HTML tree （SimHTree for short）, which reduce the time cost for the identification of data record. As the second step of the data extraction problem, substring match and adjustment algorithm is proposed, which identify the data record and data item. Experimental results show that the proposed method is effective and efficient.

关 键 词：WEB数据抽取 WEB挖掘 结构化数据 信息抽取 

分 类 号：TP393.092[自动化与计算机技术—计算机应用技术]