A Risk Prediction Model of Hard Landing Based on Random Forest Algorithm

Landing safety is a hot issue in civil aviation management. In order to fully mine the influence factors of hard landing flight data and effectively predict risk landing, random forest algorithm was introduced. Firstly, this paper qualitatively analyzed chose features model based on data. Secondly, gives quantitative analysis method importance Gini index. Finally, for dataset class-imbalanced, training SMOTE method. Then, forests classifier built verified by real The results showed that recall rate 85.50%. can not only prevent occurrence but also provide reference airlines apply mining improve ability events