An Ontology Evolution Algorithm For Reducing Impact Ranges Based On Dependency Graph Model

In computer science, ontology is the formal specification and the hierarchicaldescription of knowledge. Ontologies are often large and complex structures, whosedevelopment and maintenance give rise to certain sturdy and interesting research problems.One of the most important such problems is ontology evolution, which is the problem ofmodifying an ontology in response to a certain change in the domain or itsconceptualization. In the Semantic Web era, the changes of ontologies come to be a keyfactor for the services provided by ontology based web applications. Because of theuniversality and importance of the ontology, ontology evolution is an intrinsic part of theSemantic Web. Stojanovic has presented a procedural approach and a declarative approachto maintain ontology consistency. In the procedural approach the author proposed evolutionstrategies. Ontology engineers need to understand and participate in the process of ontologyevolution. Since the approach overly depends on humaning, so in the declarative approachthe author simplifies the changing needs, and ontology engineers can pay close attention forthe result of ontology evolution.How to reduce the impacts of an evolution procedure has became an important issue inontology evolution. After studying the semantics of the description logic SHOIN(D), weclarify the dependency relations between the elements in the evolving ontology and proposea dependency graph model for the ontology. Based on this model, the concept of impactfrequency is defined and an algorithm is proposed to compute it. We believe the impactfrequency of an ontology can be a good candidate for evaluation of the influence in theontology evolution. In this paper, we present the general representation of a change in the evolution and therepresentation in description logic SHOIN(D), then we prove that any complex ontologychange operation can be broken down to fundamental change operation. Secondly, weclarify the dependency relations between the elements in the evolving ontology and proposea dependency graph model for the ontology. Thirdly, based on this model, the concept ofimpact frequency is defined and an algorithm is proposed to compute it. Finally, we presentan ontology evolution algorithm for reducing impact ranges.Further research works include taking dependency relations between the elements in theevolving ontology and ontologies outside into account in the impact ranges, compressingthe matrix, quantifying the affected area of the entity when proving the effectiveness of theentity in the ontology.