Flight Arrival Delay Time Prediction Based on Machine Learning

Precise flight delay prediction is an important foundation to make airline business perfect. Up-to-date research have been devoted to applying machine learning and deep learning methods to predict the flight delay. The predicted task of a majority of the previous flight prediction are whether the leaving flight is delayed. The paper explores the effect of the duration of leaving delay to the arriving delay. To predict the arriving delay, a double layer model is established to complete the mission. The first layer predict the arriving delay level(0-5,in 15 minutes). With the predicted delay level, the data is input to the second layer to predict the concrete duration of the arriving delay. Compared with a single method applied to predict the duration of delay, the double layer model can contain higher prediction accuracy and efficiency.