Complex Networkbased Ontology Structure Analysis

Abstract:

Concepts and relations between concepts in ontology can be considered as the vertex and edges in network due to their similarity in structure.As a result,the analyzing method and metrics for complex network may also be used to define the complexity of ontology.In the study the degree distribution,clustering coefficient and average distance of GO ontology are measured to analyze its complexity.The result shows that the GO ontology has smallworld property but does not have scale free property.There is no clustering characteristic in GO ontology.The study can be used to evaluate importance of ontology concepts and to improve the efficiency of semantic matching.

Key words:complex network; ontology; clustering coefficient; smallworld; scale free