The Application Research Of A Non-Taxonomic Relation Extraction Method Of Ontology

Ontology learning mainly includes concept extraction, taxonomic relation extraction and non-taxonomic relation extraction. In recent years, text-based ontology learning in the concept extraction and taxonomic relation extraction advances a lot, but that in non-taxonomic relation extraction has not been extensively. Meanwhile, these non-taxonomic relation extraction methods cannot get ideal results, and all have a common defect-- only stay in the judgment that whether two concepts have a non-taxonomic relation but cannot get the tag of the relation. The tags of non-taxonomic relation are required to be manually labeled by ontology engineers. Accordingly, a further research on non-taxonomic relation extraction is needed.After analyzing the feature of OWL language and the character of non-taxonomic relation, some concepts about NNV( Noun-Noun-Verb, noun noun verb) association rule are defined based on the combination of natural language processing and association rule, and a method of non-taxonomic relation extraction based on NNV association rule is given. The method adds the construction of synonym table, extends the construction of domain dictionary, enriches the field terms filtration process based on the flow of concept extraction,gives the transaction sets acquisition algorithm and the calculation methods of NNV_support and NNV_confidence, and describes the process of non-taxonomic relation extraction detailedly. Using this method, non-taxonomic relation of Civil Aviation emergency domain ontology are extracted. Then, after the analysis and experimental comparison to other methods, the feasibility of this method is verified. The research provides support to the method of ontology learning and promotes the application research of Civil Aviation Emergency management.