Research On User Portrait Construction Method Of Civil Aviation Uncivilized Passengers

In recent years,uncivilized passengers of civil aviation frequently occur,which not only seriously affects the normal lives and interests of passengers,but also seriously affects aviation safety.How to use civil aviation uncivilized passenger data to classify uncivilized passengers according to the degree of uncivilized behavior and provide decision-making basis for airlines and airports is particularly important.User portraits are abstracted by information such as the user's basic attributes and behavioral content,"tagging" users,and classifying users.In addition,the existing research on user portraits focuses on the numerical characteristics of the data,but text as an important data characteristic has not been fully utilized.Therefore,this paper takes the textual data of civil aviation uncivilized passengers as the research content,and focuses on the research of the user portrait construction method of civil aviation uncivilized passengers.This paper first proposes a graph-based convolution and attention model for locationaware bidirectional long-term and short-term memory networks for target sentiment analysis,which provides an algorithm model and theoretical basis for constructing user portraits of civil aviation uncivilized passengers.The model is divided into two channels,which use two-way long-term and short-term memory networks and multi-head attention mechanism,locationaware two-way long-term and short-term memory networks,and graph convolutional neural networks on the dependency tree for feature extraction.Experiments,through the restaurant and laptop data of SemEval 2014 task 4,show that the model significantly improves the classification accuracy.In addition,in order to realize the classification of civil aviation uncivilized passengers,the user profile method of civil aviation uncivilized passengers based on target sentiment analysis is further studied.Using text data of civil aviation uncivilized passengers,a user portrait model of civil aviation uncivilized passengers based on target sentiment analysis is proposed.The model first extracts features through multi-head attention interaction,and then uses a gated convolution mechanism to capture features.Experiments show that the model effectively improves the accuracy of civil aviation uncivilized passenger classification and reduces the convergence time.