Research On Semantic Integration Technology Of Distributed Ontology

With the continuous development of Internet technology,there are more and more ways and types for people to acquire knowledge.Ontology can describe the knowledge in a specific field and organize the knowledge objects in this field to help relevant personnel understand the relevant elements and relationships more clearly and accurately,thus providing the basis for further processing of knowledge.As the information sources of ontology builders are distributed and autonomous,they often adopt different modeling methods or ontology description languages,but the contents described by these ontologies sometimes overlap or are related semantically,which results in the heterogeneity of ontologies.Heterogeneity of ontologies prevents direct interaction between distributed ontologies,which hinders reuse and sharing of semantic knowledgeHow to effectively fuse distributed ontologies so as to realize knowledge reuse and sharing is the main research direction of this paper.The specific research contents and methods include:(1)Research ontology aggregation algorithm based on social selection.Ontology aggregation is one of the effective ways to solve ontology heterogeneity.It aims to form a shared top-level ontology from multiple heterogeneous source ontologies constructed by agent through ontology merging mechanism,so as to form a larger semantic shared space.This study regards ontology aggregation as an application of social selection,which is used to analyze the relationship between individual source ontology and decision-sharing ontology.Firstly,the ontology merging framework and specific process are constructed based on social selection and description logic.On this basis,a distance-based ontology clustering algorithm is designed to reduce the adverse effects of untrusted ontologies on merging results.Secondly,the aggregation function in social selection is summarized and improved,and it is applied to ontology merging.The integration aggregation rule and the ladder aggregation rule are introduced.Finally,the attributes of ontology aggregation rules are analyzed,and the effectiveness of the algorithm is verified through comparative experiments(2)Study the auction-based one-to-many ontology mapping method.This content transforms the entity correlation problem into a group decision problem,formalizes the ontology mapping decision problem into a combinatorial auction problem,and proposes an ontology mapping model based on decision theory and auction theory.The ontology mapping model consists of a user input module,a correlation aggregation module,an ontology mapping decision module and a user output module.Users first input heterogeneous ontologies and enter the correlation aggregation module after normalization.In this module,the names,semantics and other aspects of ontology elements will be comprehensively considered to generate the correlation matrix between entities.Then the correlation matrix is used as input to the ontology mapping decision module,which regards the source ontology and the target ontology as buyers and sellers respectively based on auction theory and determines the optimal matching pair through auction rules.(3)Research the ontology negotiation mechanism under the agent interaction model.This research combines ontology negotiation with agent communication and designs an ontology automatic negotiation mechanism based on agent interaction model.The mechanism includes negotiation object selection protocol,agent interaction protocol and concept matching protocol.Firstly,an agent negotiation object selection method based on personal preference is proposed in the negotiation object selection protocol.This method can make agents who abide by the protocol have a greater chance to participate in ontology negotiation.In concept matching protocol,cosine similarity is used to determine the mapping between concepts Finally,in the negotiation object selection protocol,an agent strategy is designed to handle knowledge exchange between agents.In addition,it is also proved that the mechanism guarantees the lossless transmission of information.At the same time,the existing experimental results also verify the effectiveness of the mechanism.This paper studies the integration technology of distributed ontology.The results show that the reuse and sharing of knowledge can be realized,and the accuracy and recall rate of system query can be improved in the fields of knowledge integration,intelligent question answering system,information retrieval,etc.