Is the Semantic Web a Small World?

Abstract

In this paper, we describe a comprehensive analysis of graph-theoretical properties of online social networks based on the Friend-of-a-Friend (FOAF) ontology. Of particular interest for this work were properties related to the small-world phenomenon. More than 1.6 million of the FOAF documents collected on the Semantic Web met our requirements and were analyzed in depth. Most FOAF documents are created and published by social networking services, blog hosting services, or combinations of the two as a matter of routine; only a fractional amount are maintained by individuals. Although the FOAF ontology defines unique identifiers for persons in theory, retrieval and particularly fusion of personal information is difficult and error-prone in practice. Nevertheless, we identified the largest strongly connected components of various community networks based on FOAF documents and analyzed them in regard to the small-world phenomenon. Interestingly, all components examined exhibited a characteristic path length comparable to the smallest length achievable for a graph of the respective size, and the clustering coefficient was much greater than expected for an equivalent random graph; along with power law degree distributions, both are typical features of small-world graphs.