Flight Delay Prediction Based on Gradient Boosting Ensemble Models

EasyChair Preprint 8511, version 2

Abstract

In recent years, the volume of airline transportation has increased with the rapid development of aviation. With an increased demand for flights, aviation is confronted with the issue of flight delays, which becomes a series of issues that must be addressed efficiently. Correct flight delay prediction can improve airport operations efficiency and passenger travel comfort. The current study uses Gradient boosting ensemble models to build a machine learning flight delay prediction model. The Airline dataset was subjected to three different gradient boosting techniques: CatBoost, LightGBM, XGBoost, and Decision tree. To validate the performance and efficiency of the proposed method, a comparative analysis between the top performed Boosting techniques with other Ensemble Techniques is performed. CatBoost improves prediction accuracy while maintaining stability, according to the comparison results on the given dataset.

Keyphrases: CatBoost, GBoost, LightGBM, XGBoost, airline, delay prediction, machine learning

@booklet{EasyChair:8511,
  author    = {Rahemeen Khan and Salas Akbar and Tooba Zahid},
  title     = {Flight Delay Prediction Based on Gradient Boosting Ensemble Models},
  howpublished = {EasyChair Preprint 8511},
  year      = {EasyChair, 2022}}