当今互联网已成为一个巨大的开放式知识库，其中包含着许多有价值的信息。互联网信息呈现形式多样性的 特点，如何初步筛选出有价值的网页，是信息抽取的第一要务，也是构建知识库的基础。本文在建立互联网模型基础上，利用 Hadoop 平台下的Pagerank 算法，旨在研究如何在节省时间和空间基础上筛选出有价值的网页，为从互联网抽取有价值信息构 建知识库提供解决方案。

The network has become the biggest knowledge base and contains a lot of valuable information. The presentation form of Internet information is diversified. How to discover valuable page is top priority of information extraction and the foundation of building knowledge base. Based on the Internet model, this article researches how to discover valuable pages using Pagerank algorithm in Hadoop platform saving time and space, to provide solutions for knowledge base construction.