Flight Delay Prediction Using Gradient Boosting Machine Learning Classifiers

With the increasing of civil aviation business, flight delay has become a key problem in field recent years, which brought considerable economic impact to airlines and related industries. The prediction specific flights is very important for airlines’ plan, airport resource allocation, insurance company strategy personal arrangement. influence factors have high complexity non-linear relationship. different situations various regions airports, even deviation or airline arrangement all certain on delay, makes more difficult. In view limitations existing models, this paper proposes model with generalization ability corresponding machine learning classification algorithm. This fully exploits temporal spatial characteristics higher dimensions, such as preceding flights, situation departure landing overall same route. process learning, trained historical data tested latest actual data. test result shows that algorithm can provide an effective method delay.