Research On Automatic Matching Method Of Civil Aviation Uncivilized Passenger Rules Based On Deep Learning

With the vigorous development of the civil aviation industry and the continuous improvement of people’s living standards,more and more people choose to travel by civil aviation.However,in the travel process of civil aviation passengers,the uncivilized behavior of some passengers has seriously affected the safety and efficiency of civil aviation operations.How to quickly match uncivilized passengers to reasonable punishment rules is of great significance for maintaining civil aviation safety and improving civil aviation efficiency.The research on the automatic matching method of civil aviation uncivilized passenger rules is to use the existing text data to complete the matching task by extracting the features of the text and then calculating the similarity,and further use keyword extraction to remove redundant and repetitive texts to generate concise new rules.Therefore,this paper takes the text data of civil aviation passengers’ uncivilized behavior information as the research object,and focuses on text matching based on semantic similarity and keyword extraction.This paper realizes the matching of uncivilized passengers with rules and the rapid generation of concise rules,which is of great significance for the precise management of uncivilized passengers in civil aviation.First,this paper proposes a gru-capsule combined network model based on capsule network to match text similarity.The model has two channels.The two sentence vectors are extracted through the same network structure.They pass through the gated recurrent network and then the capsule network.After the feature is extracted,the similarity calculation and classification are performed.Experimental results show that compared with other models in this paper,this model has the highest accuracy on all three data sets.At the same time,in response to some uncivilized behaviors that fail to match reasonable and concise punishment rules and the redundancy problems caused by the very similarities between existing rules,through the analysis of the civil aviation uncivilized passenger data set,a method is proposed to load civil aviation stop words and reserved vocabularies and use part of speech importance to optimize the influence The Text Rank algorithm.This method first manually label stop words and reserved words,then construct a civil aviation word map,and generate a weight matrix,and finally run an improved algorithm to generate keywords.Experiments show that the keywords extracted by this method are the closest to manually labeled keywords,with the highest F value and Rouge value,and new rules can be quickly formed through the generated keywords.