A New Way of Extracting the Topic Information in Web Pages Based on DIV Tag-tree
基于DIV标签树的网页主题信息抽取方法①

Since CSS+DIV Topological Mode has become the major trend of the structural layout of web pages, the efficient extraction of the topic information in these web pages has become one of the urgent tasks for all professional surfing engines. This paper puts forward a new way of extracting the topic information in web pages based on the DIV tag-tree. It divides HTML files into DIV-forest with the help of DIV-tag. Then it filters the noise nodes in DIV tag-trees and sets up STU-DIV model-trees. Finally, it crops the DIV tag-trees irrelevant to the topic information by Topic Corelation Analysis and Cut-Tree Algorithm. It proves that this method can efficiently extract the topic information in web pages by analyzing several news web pages .