Structural Properties of Linked RDF Documents

There is an ever-growing number of diverse RDF documents available on the Web, and most of them are published following Linked Data principles. To indicate the current status of data interconnection, we analyze structural properties of these linked RDF documents. We propose a document link graph DocGraph to model links between documents, and analyze its structure from three aspects: degree distribution, morphological structure, and reachability. We report our experiments on structural properties of the graph using two crawls, each with about 10 million documents. We find that the DocGraph is scale-free, and with small average distance. Its structure in 2012 is close to that of the Hypertext Web in the years around 2001---2002, and is not as good as the structure of the Hypertext Web in later years. Therefore, we conclude that data interlinking is very necessary for the Web of Data.