Uncertainty Representation And Reasoning Research Of Ontology Based On Fuzzy MEBN

With the rapid development of the Semantic Web research,large amount of uncertain data springs up and brings great challenge to application of Semantic Web,which make representation and reasoning with uncertain information become even more important.Despite ontology is capable of modeling the semantic and knowledge in knowledge-based system,classical ontology languages are not appropriate to deal with uncertainty in knowledge,which lead to the need for extension of the ontology language.At present,main extension methods only focus on either probability or fuzziness,while they always appear together in the real world application domains.In view of this situation,this thesis analyzes and summarizes the research of uncertainty representation and reasoning,then undertakes the study of fuzzy-probability ontology and proposes an ontology representation and reasoning frameworks based on Fuzzy Multi-Entity Bayesian Networks(Fuzzy MEBN).The article mainly includes the following works:First of all,we research on the representation approach of uncertain knowledge in the Semantic Web.Then we combine the Fuzzy MEBN and ontology to propose a Fuzzy Multi Entity Bayesian Networks ontology language based on OWL2(Web Ontology Language)which can construct the classes and properties of fuzzy probability knowledge.The thesis defines and studying its syntax and semantics,and showing representation of domain knowledge by examples.Next,we study the reasoning approach of uncertain knowledge and proposed a reasoning framework based on the representation of FuzzyPR-OWL,we combine the fuzzy probability theory and belief propagation algorithm to propose a fuzzy probability-based belief propagation algorithm which add fuzzy factor to the belief that propagate between nodes.In addition,we introduce the process of reasoning in FuzzyPR-OWL frameworks,include SSFBN structure algorithm of fuzzy MEBN,data fuzziness,inference of fuzzy rules and fuzzy belief propagation.Finally,the feasibility and validity of the representation and reasoning method of fuzzy probability ontology are proved by the experiment.We construct domain ontology of Collision Warning System using proposed FuzzyPR-OWL and calculate the probability of target variable.We make comparisons of the result affected by fuzzy states and the result which haven't been affected,then use the ten-fold cross-validation method to evaluate the accuracy of the algorithm.The result of experiment shows that proposed frameworks can model and reasoning fuzzy probability knowledge properly.This thesis provides a new solution for the uncertain reasoning and representation.