Research On Construction And Semantic Retrieval Of Multiple Majors Domain Ontology

Ontology is playing an increasingly important role in the fields of software engineering, artificial intelligence, information retrieval and web service research. According to the level of dependence on research fields ontology can be divided into Top Ontology, Domain Ontology, Task Ontology and Application Ontology. Domain ontology can effectively organize the knowledge of that domain and make it easier to share and reuse. Some kinds of domains such as agriculture, railway, high-speed railway and aviation include different major fields. For example, the high-speed railway domain consists of maintenance engineering, traction power supply, EMU and operation management etc. major fields. Oriented to the domain which consists of several major fields, this dissertation researches construction methodology for multiple majors domain ontology and semantic retrieval and reasoning based on multiple majors domain ontology. This dissertation makes following contributions:(1) Proposes MMDOB (Multiple Majors Domain Ontology Building) methodBased on the construction methods of domain ontology and the feature of multiplemajors domain knowledge, and oriented to the domain which consists of several major fields, this dissertation puts forward MMDOB method based on thesaurus and thematic words. First, MMDOB method builds ontologies of the major fields based on thesaurus and thematic words, and then integrates these ontologies into a unified ontology for the multiple major domain. According to actual situation of the domain which consists of several major fields, this dissertation proposes MMDOI (Multiple Majors Domain Ontology Integration) method and three layers ontology integration framework for multiple majors domain.(2) Gives a model for multiple majors domain ontology and the method for concept semantic similarity computationBased on the ontology model and multiple majors domain ontology, this dissertation proposes Eight-tuples model for multiple majors domain ontology, Nine-tuples model for the concept of multiple majors domain ontology and formalized description for multiple majors domain ontology.Concept semantic similarity computation is an essential method of semantic extension retrieval. On the basis of concept model for multiple majors domain ontology, this dissertation builds MD4 (Fourfold Matching-Distance Model) to compute concept semantic similarity and it also provides the computation detail.(3) Gives five kinds of semantic expansion modelsExtended retrieval is through searching the related concepts in search terms to find the related knowledge. According to the feature of multiple majors domain ontology and the search object, this dissertation analyses semantic expansion model of user query and proposes five models for semantic expansion retrieval. This dissertation chooses description logic that is the logic foundation of OWL as the reasoning foundation.(4) Gives an empirical study for High-speed railway domainUnder the guidance of MMDOB method, this dissertation builds various majors ontology of high-speed railway, and they are preliminary merged into one unified domain ontology of high-speed railway. On this basis, this dissertation designs and develops HSRK-SRRS (High-speed Railway Knowledge-Semantic Retrieval and Reasoning System). It tests MMDOB method, semantic extension retrieval and reasoning model.MMDOB method that is proposed in this dissertation has some practical significance, and it also has a certain reference for ontology construction of multiple majors domain, semantic relations analysis between concepts and ontology integration of multiple majors domain. This dissertation builds MD4 based on the model for multiple majors domain ontology to compute domain ontology Concept Semantic Similarity. The experiment to test MD4 shows that the model can reflect the semantic relation between concepts. So it gives an effective method to quantify the semantic relation between concepts in domain ontology that are constituted with several major fields. These research results are applied in high-speed railway domain, and that provides a positive reference for the acchivement of ontology construction and semantic retrieval and reasoning for the domain which consists of several major fields.