CrimAr: A Criminal Arabic Ontology for a Benchmark Based Evaluation

Recently, ontologies have become more important in modern Semantic Web as they capture knowledge in a particular domain of interest. Indeed, they emphasize interoperability and establish a common shared understanding among the involved actors of web-based applications. Nevertheless, in parallel with the abundance of the proposed approaches for ontology learning, a related problem of the evaluation of such automatically generated ontologies is emerging in different domains. In the Arabic legal domain, a benchmark golden ontology is so necessary in order to assess the good quality of the (semi-)automatic learned ontologies. In this paper, we introduce CrimAr, a handcrafted ontology based on the top-levels of LRI-Core, to represent all relevant knowledge in the Arabic legal domain, especially the criminal matter. The use of CrimAr is also demonstrated in a real case evaluation.