A novel multi-aspect consistency measurement for ontologies

Web developers have started to integrate semantic information to their systems increasingly often. The semantic metadata embedded with the resources is typically linked to ontologies or taxonomies. Meta information can bring a number of advantages for users. However, the ontologies might contain some errors or could be partially inconsistent. Therefore it is important to evaluate the quality of ontologies at various levels. Existing evaluation methods either investigate whether the ontologies are "fit for purpose", or focus on evaluating ontology consistencies from a single aspect. In this study, we focus on ontology consistency evaluation methods, which consider lexical, taxonomic and syntactic aspects at the same time. We propose new measures, which capture several essential aspects simultaneously. We demonstrate the effectiveness of the new measure through a case study and an extensive set of experiments.