基于树形结构的Web信息抽取

Web Information Extraction Based on Tree Structure

提出了一种基于树形结构的Web结构化数据抽取算法.该算法基于HTML的树形层次结构,包括HTML树构造算法,数据区域挖掘算法,数据记录挖掘算法以及数据记录模式生成算法.算法引入了页面元素布局位置等信息用于清洗页面,采用层次划分思想实现页面数据区域的挖掘,并通过树匹配生成记录模式,实现最终数据项抽取.实验表明,该方法可以有效地实现Web结构化数据抽取.

It proposes tree structure based Web data extraction algorithm in view of the inadequacies of the existing methods. The tree structure based algorithm includes： the algorithm of HTML tree construction, the algorithm of data region mining, the algorithm of data record mining, and the algorithm of record schema generation. The algorithm cleans the Web pages using the position information of page elements, mines data region by hierarchical Clustering, and generates record schema finishing data item extraction through tree matching. Experimental results show that our algorithm can-improve the accuracy and efficiency of Web data extraction.