Research on Taxiout Time Prediction of Departing Aircraft Based on Stacking Ensemble Learning

2026-99-2423

11/20/2026

Authors
Abstract
Content
To improve the accuracy of taxiing time prediction for departing aircraft at hub airports, this paper constructs a taxiing time prediction model for departing aircraft using BP neural networks, random forests, and XGBoost as base learners and introducing the Stacking ensemble learning framework. Key factors affecting taxiing time are analyzed and categorized into core feature groups and all feature groups through correlation testing. Taking Guangzhou Baiyun International Airport as a case study, the prediction performance of each model is evaluated using performance metrics such as root mean square error (RMSE), mean absolute error (MAE), and coefficient of determination (R2). Experimental results show that introducing all feature groups effectively improves the model’s prediction accuracy compared to using core feature groups. Compared with models using BP neural networks, random forests, and XGBoost, the prediction model built based on the Stacking ensemble learning method exhibits smaller prediction errors, validating the effectiveness of the method.
Meta TagsDetails
DOI
https://doi.org/10.4271/2026-99-2423
Citation
Tao, S. and Kou, W., "Research on Taxiout Time Prediction of Departing Aircraft Based on Stacking Ensemble Learning," 2026 2nd International Conference on Intelligent Transportation and Future Mobility (ITFM2026), Xi'An, China, March 27, 2026, https://doi.org/10.4271/2026-99-2423.
Additional Details
Publisher
Published
Yesterday
Product Code
2026-99-2423
Content Type
Technical Paper
Language
English