A Delay Prediction Method for the Whole Process of Transit Flight

Abstract

In order to strengthen the construction of smart airports and improve the ability of airport managers to identify, intervene and rescue delayed flights, this paper proposes a delay prediction method for the whole process of transit flights through the basic steps of node time and link time prediction and delayed flight identification. By designing the key node time prediction model (ML-DM), the method predicts the important guaranteed node time involved in the process of flight departure from the outstation to the departure from the current station. By constructing the imbalance data classification model, the delayed flight is identified at each predicted guarantee node. The experimental results for a busy airport show that this prediction method can achieve a maximum recognition rate of 96.5% for delayed flights.

Keywords

• transit flight
• delay prediction
• ML-DM model
• imbalanced data classification model