Selection Function Optimization Strategy For Ontology Debugging Of Black-Box

In the ontology applications,it is very important to build a high quality ontology and manage it efficiently.However,in practical applications,it is difficult to build an ontology without errors.The reasons for the inconsistency of ontology are mainly the wrong understanding and definition of the concept,the polysemy,the change of ontology format,and the merging of different sources.Under the classical reasoning,if the ontology contains logical conflict,it can deduce any conclusion.So it is meaningless to reasoning the ontology containing the logical conflict.In order to avoid this situation,it is necessary to diagnose the logical conflict of the ontology,which is the most important thing to ensure the quality of the ontology and improve the efficiency of reasoning.With introducing the research status of ontology debugging,and based on the analysis of the black box and white box of ontology debugging method,the selection function algorithm of black box is mainly studied.The main contributions of this paper include:(1)The depth first search strategy is applied to the selection function algorithm.The algorithm is based on the axiomatic relevance axiom in the expansion stage of ontology debugging,and depth first search strategy is carried out in the process of selecting axioms in the axiom graph.The algorithm is realized by the stack.(2)The caching window mechanism is proposed to record the number of axioms that the selection function searches.In the selection function algorithm,the reasoning machine is invoked once every new axiom is added,which causes the waste of time and resources.To solve this problem,the caching window mechanism is proposed inthis paper,which is to set up a caching window in the expansion process.Only when the number of new added axioms increases to the size of the caching window,can we invoke the reasoning machine.(3)With the proposed depth first search strategy as framework,the heuristic algorithm of selection function is proposed based on the characteristics of conjunction operator and disjunction operator.In the description logic language(32)(43)(34),a complex concept is mainly composed of conjunctive operator and disjunctive operator and atomic concepts and atomic roles together.Based on the fact that the influence of the conjunction operator and disjunction operator on the unsatisfaction of concept is different,the heuristic depth first search algorithm is proposed,which encountering complex concept in the expansion process,give priority to expansion of the concept of composite containing conjunction operator axiom.Through the experiment comparison and analysis,it is concluded that the depth first search strategy proposed in this paper is effective for ontology inconsistent debugging.For the ontologies with more axioms,the caching window mechanism can improve the efficiency of the selection function algorithm.And when the number of disjunctive operators in ontology is large,the heuristic selection function algorithm has high efficiency.