•  
  •  
 

Journal of System Simulation

Abstract

Abstract: Aiming at the classification of flight delay under imbalance data, a novel method based on nonlinear weighted XGBoost (extreme gradient boosting) is proposed. The imbalance of flight delay data and the influence for classification performance caused by the data imbalance are analyzed. A heuristic nonlinear weighting method based on sample proportion is proposed, and the negative log likelihood loss function is optimized. The real flight delay dataset is used to validate the performance of the classification algorithm. The experiment results show that the proposed nonlinear weighted XGBoost algorithm can improve the classification accuracy of flight delay, while ensuing a high overall classification accuracy. Compared to traditional methods, the proposed algorithm has good performance of statistical metrics and performance curves.

First Page

2261

Revised Date

2020-08-05

Last Page

2269

CLC

U8;TP391.9

Recommended Citation

Tang Hong, Wang Dong, Song Bo, Chu Wenkui, He Linyuan. Classification of Flight Delay Based on Nonlinear Weighted XGBoost[J]. Journal of System Simulation, 2021, 33(9): 2261-2269.

DOI

10.16182/j.issn1004731x.joss.20-0372

Share

COinS