Ontology Metrics in the Context of the GF Framework for OBDA

Abstract

There is a need of generating high-quality ontologies for Semantic Web applications that need to access legacy, relational and nonrelational data sources. We present GF, a tool for materialization of ontologies from relational and non-relational data sources such as H2 databases, CSV files and Excel spreadsheets. We evaluate a sample case generated by GF with a third-party ontology evaluation tool called OntoMetrics. The results obtained show that the ontologies generated with GF are reasonably good for being used in Semantic Web applications as they are validated correctly and pass all of the filters for the OWL2 main profiles, thus making them suitable for processing with lightweight reasoners. The metrics indicate that our application is lacking quality in the annotation area regarding the documentation of the classes and properties generated by the application.