Flight delay prediction has nonlinear aggregation dynamics.In order to improve the pre-diction efficiency under the premise of ensuring the accuracy,a LightGBM(light gradient boosting machine)algorithm based on airport aggregate departure delay prediction model is proposed.Through the analysis and processing of historical flight data,three important features,the time feature,the flight plan feature and the delay feature,are extracted,and the extracted characteris-tics are taken as input variables.The LightGBM algorithm is used to predict the flight delay time based on the historical operating data of Guangzhou Baiyun Airport(ZGGG).The results show that the predicted delay time is in good agreement with the actual delay time.Compared with the prediction results of other commonly used algorithms,the proposed model has better effect and higher efficiency in various prediction indexes.
维普
万方数据