Towards Measurement, Evolution And Modeling On China Web Graph

The Web pages and hyperlinks among them induced a large scale directed graph, Web Graph. The huge graph contains rich link information and interesting properties such as the power law degree distribution, bowtie structure and etc. Expoliting the link information behind the Web Graph has great impact on resource managing and Web application development.China Web Graph is an important sub-graph of the global Web Graph. It contains much localization properties. This makes its properties different from those of global Web Graph.In this thesis, we systematically analyze the huge link data crawled by SkyNet in April 2004 to study the static properties of China Web Graph. Meanwhile, by the observation of 150 various Web sites in China, the evolution of China Web Graph is studied. Finally, on the basis of the research of properties and evolution of China Web Graph, a hierarchical model of Web Graph is proposed to extend the state-of-the-art modeling methods of Web Graph.