Incentive Factors Analysis And Risk Evaluation Of Aircraft Approach And Landing Based On Text Mining

Flight safety is an eternal issue discussed during the operation of the civil aviation system,the approach and landing flight phase with the highest probability of flight accidents.In this stage,the altitude of the aircraft is low,the speed is small,the time is short,and the operation tasks are complicated.Once an emergency occurs,it is easy to induce the occurrence of flight accidents.Therefore,the research on the key risk factors and risk control of flight approach and landing has important practical significance to ensure the safety.This paper takes the whole process of the approach and landing stage as the research object,uses 11277 unsafe event information texts in the flight approach and landing stage as the basis,and combines text analysis methods,neural network models,risk analysis theory,and front-end and back-end development technologies to realize text-based mining.Research on risk factors analysis and risk assessment of aircraft approach and landing.Main research contents and results:(1)Extraction of key risk factors for approach and landing.Using text mining method,through text preprocessing,feature item extraction,dimensionality reduction and construction of space vector model,30 dimensionality reduction results of approach and landing features are formed,and the optimized K-means algorithm is used to cluster each risk factor set.Class Analysis and Induction.(2)Construction of multi-channel Text CNN-Bi GRU risk factor classification model based on attention mechanism.The neural network model is used to classify the text of risk factors,so as to quickly find the key factors and their theoretical weights.(3)Analysis of the importance of key risk factors.The importance of risk factors is analyzed by combining the theoretical weights of key risk factors and the grey relational method,and the traditional triangular fuzzy accident tree method is used for verification.(4)Construction of an approach and landing accident risk factor identification system.The research content is integrated,and the Django framework is used to initially form the function of approach and landing risk factor search and data visualization.The research provides basic guarantees for finding potential safety hazards,preventive control,safety capability during approach and landing,and establishment of risk management systems.In this paper,key risk factors including manipulation errors,communication errors,navigation precision errors,and unreasonable management procedures are obtained through text analysis.The grey correlation method is used to evaluate the risk factors of risk factors,and a dynamic identification system for the risk factors of approach and landing accidents based on information text analysis is designed and developed,which provides a reference for the approach and landing risk management and control based on information data-driven.