基于梯度提升决策树的航班延误分类预测

Classification of Flight Delay Based-on GBDT

考虑航空交通网络状态特征对航班延误的影响，将上游的航班延误状态特征加入到预测因素中，并使用梯度提升决策树（Gradient Boost Decision Tree，GBDT）的方法构建了航班延误预测模型．与以往的决策树算法、SVM分类算法、RF算法相比，GBDT算法在航班延误分类预测上具有更高的准确度，可有效提高机场运行管理效率．

Consideing the impact of flight delays by characteristics of air transport network, the article added the features of flight delays to the prediction, and used Gradient Boost Decision Tree, （GBDT）to construct model for prediction of flight delays. Decision tree algo- rthm, compared with past algorithms of SVM, RF algorithms than GBDT algorithm in the prediction of flight delays with higher accuracy, which can effectively improve the efficiency of airport operation and management.