THE EVALUATIONS FOR THE BACKEND OF ONTI MEASURES WITH BLACK BOX METHOD

Inconsistency in an ontology can be a serious problem since it can mess up the information in the ontology. Ontology-based inconsistency measure gives inconsistency value of the whole base of the OWL ontology. It means the produced inconsistency value is used to evaluate its whole base. Based on this characteristic, there were 10 inconsistency measures created in the previous research and collected into one package of measures in an application program, namely Onti Measures. The application will not be useful if the measures do not work well. This problem leads to conduct evaluations. In this research, evaluations for the backend part of Onti Measures with the use of three kinds of OWL reasoners are done to know the performance of the application system with the comparison of each reasoner usage. The evaluations for the whole part of the application are not the scope of this research since they are only done for the backend part. Particularly, they are done with the black box method since the structure of the codes are not necessary to be known. They are evaluated with several OWL files as test cases and as the inputs of the backend program. The evaluation shows that the same inconsistent OWL file that is computed with a different type of inconsistency measure with any chosen reasoner may result in different inconsistency value. Other evaluations are provided. Overall, they show that Pellet is better than the two other reasoners and I_(D_f ) is more efficient than the other measures.