A CNN-LSTM Framework for Flight Delay Prediction

Abstract

In this work, we proposed a CNN-LSTM deep learning framework to predict flight delays. The proposed CNN-LSTM model consists of three main components: A Convolution neural network (CNN) followed by a Long short-term memory (LSTM) network, and then a Random Forest classifier is applied for flight delay prediction. First, a convolution neural network framework is employed to extract the spatial corrections among different regions. Then, the results of the CNN algorithm are inputted into the LSTM network for temporal dynamics modeling. Finally, we use the spatial-temporal correlations obtained from the CNN-LSTM framework and extrinsic features (e.g., airline issues, distance, schedule fly time), and predict future delays with the Random Forest classifier. The U.S. domestic flights in 2019 are collected from the Bureau of Transport Statistics to confirm the outperformance of our proposed model and the result indicates that the CNN-LSTM model exhibits better performance compared with several benchmark models. The created prediction model of this study could provide useful information for airport regulators in understanding the potential delays in advance and developing effective airport management strategies for improving airport on-time performance.

Keywords: Flight Delay Prediction, Deep Learning, Convolution Neural Network, Long short-term memory (LSTM), Random Forest, Spatial-temporal Correlations