Class Association Structure Derived From Linked Objects

The Web is being extended with more and more RDF data, especially the links between objects. Object links are critical to the semantic Web just as page links to the hypertext Web. Studying the macroscopic properties of object links can help people understand the semantic Web and build (semantic) Web applications in a better manner. To achieve this, we propose a notion of class association graph (CAG) based on the object links on the semantic Web, and report the results of analyzing the complex network characteristics of a CAG constructed from a real large data set collected by the Falcons search engine. The CAG observed has the scale-free nature and small-world characteristics. Then, vertex-importance graph synopsis approach is employed to depict a landscape of such class association structure.