Research on Flight Delay Prediction Based on Horizontal and Vertical Federated Learning Framework

This paper aims to solve the low prediction accuracy caused by insufficient data samples of a single airport and the problem of data island caused by data privacy security in civil aviation. According to the characteristics of the federated learning framework, a flight delay prediction model based on federated learning is designed by integrating horizontal and vertical federated frameworks. On the premise that the original data is not out of the local area, the sample number and characteristic dimension are expanded, and an improved SecureBoost model for federated learning based on XgBoost. It is used to ensure the security of model parameters during training. The simulation of real flight control data is simulated, and the effect is compared in different machine learning models. The model proposed in this paper is compared with the traditional model, and the results show that our scheme on the premise of guarantee data privacy can meet the requirements of forecast accuracy, which proves the feasibility of our model.