Measuring Ontology Inconsistency Based on Dempster-Shafer Theory

As the size of ontologies grows and applications become more complex, it is often difficult to construct an ontology which is error-free. Inconsistencies arise naturally in the design and development of ontologies. Measuring ontology inconsistency is the basis and prerequisite of inconsistency-dealing, which can help us to decide how to act on an inconsistency, whether to ignore it or to resolve it. A number of proposals have been made for measuring the degree of inconsistency of an ontology. Most of these efforts start with flat ontology. However, the importance of ontology items is regarded unequal in numerous applications, while weights are associated with items of ontology. Therefore, measuring ontology inconsistency with weighted formulae is a significant but hard work. After analyzing the characteristics of uncertainty reasoning method based on evidence theory and ontology inconsistent measurement, this paper proposes an ontology inconsistency measurement method named ETOICM and then proves its correctness. Finally, measurement formula and algorithm are given. Compared with the related methods, features of the new method are illustrated. In addition, the measurement results of ETOICM method can be used as weights and provide basis for further resolving inconsistency and reasoning with inconsistent ontology.