Research And Application On Relation Extraction Method For Domain Ontology Based On LSTM

Domain ontology learning is a learning method that automatically acquires concepts and concepts in the domain ontology for a specific domain.With the rapid development of civil aviation industry and Internet information technology,the handling and management of civil aviation emergencies are becoming more complicated.Automatic learning of domain ontology has become a problem that needs to be solved urgently.Automatic learning of domain ontology includes automatic learning of the concepts and the relationship between concepts.This paper conducts relevant research on the automatic learning of relations in domain ontology.This paper makes a detailed study on the relation extraction method in current text information processing,and focuses on the current application status of existing deep learning methods in relation extraction,and compares the advantages of these relation extraction methods.A method of extracting the semantic relationship combine the LSTM and the attention mechanism is proposed.Firstly,the model embedding the text information and then obtain the local feature.Secondly,the local feature is introduced into the bidirectional LSTM model,and the attention mechanism is used to calculate the importance probability between the input and output of the LSTM model and then obtain the global feature.Finally,the local feature and the global feature are fused,it obtain the result of relation extraction by classifier.The experimental results on the SemEval-2010 Task 8 are superior to other deep learning methods.Using information related to civil aviation emergencies among cross-media information such as Sina Weibo,WeChat,and aviation safety voluntary reporting systems as sources of data,these textual information are highlighted,entities and relationships are marked.Combing with the above research,a domain ontology relationship extraction method based on unstructured text information is proposed.The F value of the relation extraction is higher than the existing domain ontology relationship extraction method.