To improve the efficiency of airport surface operation,it is necessary to accurately and efficiently predict the taxi time of departure flight.The influencing factors of departure flight's taxi time were analyzed and the corresponding parameters were defined,and the correlation of numerical features was analyzed.A method using Bayesian optimization(BO)to obtain the opti-mal hyper-parameters combination of light gradient boosting machine(LightGBM)algorithm was constructed for the problem that the standard LightGBM algorithm with numerous hyper-parameters lowers the prediction accuracy of the model due to the artificially set hyper-parameters.To verify the validity of the proposed model,the simulation was carried out according to the actual operation data of a large hub airport in central China,and the prediction results were compared with those of support vec-tor regression(SVR)model and BP neural network.The results indicate that the prediction accuracy and model evaluation met-rics of LightGBM with hyperparameter tuning using Bayesian optimization(BO-LightGBM)algorithm are superior to those of other methods.
关键词
航空运输/机场场面运行效率/滑行时间/相关性分析/轻量级梯度提升机/超参数优化/贝叶斯优化
Key words
air transportation/airport surface operation efficiency/taxi time/correlation analysis/LightGBM/hyperparameter optimization/Bayesian optimization
维普
万方数据