Airport Delay Prediction Method based on Improved Weather Impacted Traffic Index

To reduce the negative impact on flight delay of weather, airport delay prediction based on the improved Weather Impacted Traffic Index (WITI) model is proposed in this paper. Gaussian Mixture Model and OTSU method are used to improve the weather influence coefficient in the WITI model firstly. Correlation analysis is used to compare the original and improved WITI model. The correlation between the improved WITI model and the airport departure delay flight and delay time is 58.68% and 26.72% higher than original model. Then airport delay prediction model based on improved WITI through the random forest prediction model is constructed. Data of Shanghai Pudong International Airport (ZSPD) in 2016 is used for experimental verification, and R-square score of airport delay prediction model based on improved WITI model is 91.67% higher than the model based on original WITI model and mean square error is reduced by 84.01%.