×

Social ontologies as generalized nearly acyclic directed graphs: a quantitative graph model of social tagging. (English) Zbl 1258.68151

Dehmer, Matthias (ed.) et al., Towards an information theory of complex networks. Statistical methods and applications. Boston, MA: Birkhäuser (ISBN 978-0-8176-4903-6/hbk; 978-0-8176-4904-3/ebook). 259-319 (2011).
Summary: We introduce a quantitative graph model of social ontologies as exemplified by the category system of Wikipedia. This is done to contrast structure formation in distributed cognition with classification schemes (by example of the DDC and MeSH), formal ontologies (by example of OpenCyc and SUMO), and terminological ontologies (as exemplified by WordNet). Our basic findings are that social ontologies have a characteristic topology that clearly separates them from other types of ontologies. In this context, we introduce the notion of a Zipfian bipartivity to analyze the relationship of categories and categorized units in distributed cognition.
For the entire collection see [Zbl 1227.94001].

MSC:

68T30 Knowledge representation
05C90 Applications of graph theory
68T50 Natural language processing
91D30 Social networks; opinion dynamics
05C75 Structural characterization of families of graphs

Software:

WordNet; GermaNet; YAGO; Cyc
PDFBibTeX XMLCite
Full Text: DOI