【作者】 刘世杰； 唐世渭； 杨冬青； 王腾蛟； 李立宇；

【机构】 北京大学视觉与听觉信息处理国家重点实验； 北京大学计算机科学技术系；

【摘要】 <正>1.引言随着Internet的快速发展,Web上的信息量爆炸式地高速膨胀;随着数据量的激增,Web规模的快速增长和对网上信息的应用需求的不断提高,原有的对网页文件的链接浏览和关键词检索已经无法满足信息获取的需求。用户虽然可以得到海量数据,更多还原

【Abstract】 Web information is expanding quickly with the dramatical growth of Internet.Usually,accessible Web information which users are interested in is in HTML documents,not in databases.Those documents are always unstructured or semi-structrued,and lack of patterns and metadata.So it is very discommodious for information integation. information exchange,Web knowledge discovery,exact Web information query etc.This paper approaches a new Web information extraction and integration methology based on XML.It has four phases.The first phase creates an XML schema for each class of HTML documents,and maps it to a database schema;the second extracts patterns from the sample HTML documents,and makes a template for them;the third selects the exact template,and extracts useful contents from HTML documents according to it;the last integrates the contents into database.This approach has been implemented in the COMMIX.In addition,it achieves good effectiveness by considering universality and precision.更多还原