A novel intelligent approach for flight delay prediction

Journal of Big Data

Table 4 Accuracy of forecast horizons 6 h

Models	Traditional training model			Proposed FDPP-ML			Error reduction percent by FDPP-ML
Models	MAE	MSE	RMSE	MAE	MSE	RMSE	MAE (%)	MSE (%)	RMSE (%)
CATR	29.21	3943	62.79	22.63	2841	53.30	23	28	15
GBR	29.43	4013	63.35	23.31	3115	55.81	21	22	12
GRU	29.44	4034	63.52	23.41	3266	57.15	21	19	10
LGR	29.35	3963	62.95	22.67	2913	53.97	23	26	14
LR	29.52	4082	63.89	26.78	3735	61.11	9	9	4
LSTM	29.61	4072	63.81	24.14	3448	58.72	18	15	8
RFR	30.29	4098	64.02	22.76	2830	53.20	25	31	17
RNN	29.58	4078	63.86	23.57	3184	56.43	20	22	12
Stacking	28.83	3977	63.07	21.52	2855	53.43	25	28	15
Voting	29.38	3981	63.10	22.50	2897	53.83	23	27	15
Average error reduction percent by FDPP-ML							21	23	12