Time-Series Prediction of Intense Wind Shear Using Machine Learning Algorithms: A Case Study of Hong Kong International Airport

Machine learning algorithms are applied to predict intense wind shear from the Doppler LiDAR data located at Hong Kong International Airport. Forecasting in vicinity of airport runways is vital order make intelligent management and timely flight operation decisions. To time series shear, Bayesian optimized machine models such as adaptive boosting, light gradient boosting machine, categorical extreme random forest, natural developed this study. The time-series prediction describes a model that predicts future values based on past values. Based testing set, optimized-Extreme Gradient Boosting (XGBoost) outperformed other terms mean absolute error (1.764), squared (5.611), root (2.368), R-Square (0.859). Afterwards, XGBoost interpreted using SHapley Additive exPlanations (SHAP) method. XGBoost-based importance SHAP method reveal month year encounter location most were influential features. August more likely have high number wind-shear events. majority events occurred runway within one nautical mile departure end runway.