Research On Mechanism Of Semantic Association Based On Ontology Of Petroleum Domain

With the development of information technology, the information technology degree ofoil field production activities continues to deepen. But because of the many activities involvedin the petroleum domain, professional production process will generate the correspondingdata. These data have different resource structures, complex content and huge amount, so thatwhen the users query information, and often unable to require the data resource which theywant easily and quickly.In order to solve these problems, this thesis introduces the domain ontology theory intothe data integration management of petroleum enterprise, through in-depth analysis of theontology, ontology reasoning and other related technologies. And it presents a semanticassociation mechanism based on ontology of petroleum domain.The main contents of the thesis are as follows:Firstly, it analyzes the existing method of ontology construction of oil field through thestudy of domain ontology related technology. Meanwhile, it studies the method of ontologyconstruction of petroleum with the Seven-Step method and the definition of entities andrelationships in Epicentre data model.Secondly, according to the problem that is lack of the unified semantic standarddescriptions above all kinds of data in oil field, it researches the methods on the data resourcefeature extractions. Through the analysis of the usual data source features in production andactivities, it improves the traditional feature extraction algorithm based on TF-IDF with usingthe conceptual similarity and co-occurrence in ontology. Moreover, it accomplishes thefeature extraction about semantic data resource. Meanwhile, it puts forward the standardwhich can describe the semantic of the heterogeneous data resource.Thirdly, it analyzes the structure and function of ontology reasoning machine by thestudy of the structural characteristics of petroleum ontology. And it designs an ontologyreasoning process based on the traversal process of directed graph. In addition, it proposes thedata extraction method based on the concept relationships through the correspondencebetween ontology concepts of petroleum domain and the database dataitems of theexploration database. Consequently, it realizes the query expansion based on the reasoning ofthe ontology concept relations.Finally, the thesis tests some examples of above studies to verify the validity and feasibility of each part of the study content.