| Big data technology is constantly developing today,and flight data is also used by all the airlines company in the world.As the major employee of the airline company,the pilot,can collected a large amount of data by the QAR during takeoff,cruise,and landing,in each flight.These data Reflected the operating environment of the flight,such as meteorological,terrain and the pilot’s control.Based on the analysis of these data,targeted training and performance evaluation of pilots can effectively avoid risks in flight operations.The paper first introduced the current training system of H Aviation,and then used the Pearson correlation analysis and One-way analysis of variance to analyze the results of the questionnaire by collecting the training requirement of H Aviation pilots,and found the training requirement of pilots in training.combined with the training requirement of pilots,and the training program which is required by the CAAC and the ICAO.Differentiated theoretical training and systematic QAR training were proposed.Using the flight data of 18 sets of flight warnings,“ the distance from 50 feet to touchdown ”(code 704),the correlation analysis was performed using SPSS software,and the chi-square test was used to select the target factors.The principal component analysis was used to establish a risk assessment model,and the main reasons affecting the 704 warning were that the aircraft altitude when going through runway threshold is too high and too earlier to increase pitch attitude.Propose risk treatment measures for the conclusions and design a pilot QAR personalized training course based on the results of data analysis on the simulator course.From the content of the current flight warning evaluation,the detailed rules of the flight warning evaluation were detailed.According to the importance of various warnings,the quantitative standard weights were calculated through the superior order comparison method and the analytic hierarchy method.Finally,the performance evaluation of QAR warning was determined. |