Application Research On The Method Of Semantic Query Expansion Based On Domain Ontology

The civil aviation emergency ontology was constructed based on the emergency rescue plan, management regulations, rescue preplans, historical cases related with civil aviation It clearly describes the semantic relationships among concepts through the formal explanation of concepts and relationships. The semantic retrieval method that was previously proposed mainly considers the semantic distance between concepts on classification relationships,effectively solving the problems existing in the traditional query method, such as semantic deviation and semantic ambiguity. But there are still some problems in the query rate and precision. Therefore, a method of semantic query expansion which comprehensively considers the taxonomic and non-taxonomic between concepts is putted forward. The main research contents are as follows:1) According to the classification relationships between concepts in the constructed domain ontology, the most commonly calculation methods of semantic similarity on semantic distance, depth, semantic coincidence degree is improved by introducing the weights of relationships and the difference of depths. A new calculation method of semantic similarity is proposed based on the classification relationships. Combined with the similarity of the emergency cases, the validity of the method is verified.2) For the non-taxonomic relationships of domain ontology, a calculation method of semantic relevance is presented. This method implements the effective integration of the type and value of data property. It also resolves the calculation problem of multiple object properties between query word and concept. Finally it realizes the semantic query expansion.3) Combined with the semantic similarity based on classification relationships and the semantic relevance based on non-taxonomic relationships, a method of semantic query expansion in domain ontology is proposed. It uses different calculation method on the basis of the existence of object property between query word and concept. This method is applied into the semantic query of similar cases. According to the experiments, the accident reasons, the accident influences, rescue preplan and emergency rescue disposal measure of similar cases are successfully acquired. In the field of recall and precision, the method obtains prefer results,providing more efficient auxiliary information for the emergency decision of emergencies.