Connectivity inferences over the web for the analysis of semantic networks

Abstract

The World Wide Web constitutes a large information space and a promising support for the development of data inference, analysis and mining mechanisms. The research presented in this paper introduces a modelling approach that derives a social network and computes some of its emerging properties, from the semantics exhibited from a domain knowledge embedded in a series of web pages. The framework is applied to the research communities related to a series of conferences in a given domain, and the university connections materialised by the researcher trajectories over time. The properties of the social network are analysed using graph-based measures, and a novel geographical dispersion coefficient.