Optimization of air traffic management efficiency based on deep learning enriched by the long short-term memory (LSTM) and extreme learning machine (ELM)

Journal of Big Data

Table 6 Accuracy criteria for 15- and 30-min delays

Scenario	Delay/airport	ATL	LAX	ORD	DFW	DEN	JFK	SFO	CLT	LAS	PHX	Average
Scenario 1 80%-train 20%-test	Delay 15	98.46	88.34	88.34	98.22	98.64	98.48	90.27	96.96	94.24	97.32	94.92
Scenario 1 80%-train 20%-test	Delay 30	99.2	92.41	98.62	98.6	98.03	98.68	93.45	93.45	95.75	96.6	96.27
Scenario 2 60%-train 40%-test	Delay 15	49.99	49.98	98.48	98.36	97.83	97.28	97.28	93.51	49.92	86	79.07
Scenario 2 60%-train 40%-test	Delay 30	49.99	61.97	98.42	98.81	49.98	50	93.49	50	79.71	5.96	72.84
Scenario 3 70%-train 30%-test	Delay 15	98.71	50.34	98.28	98.64	49.99	98.31	92.22	94.99	92.74	0.8	85.42
Scenario 3 70%-train 30%-test	Delay 30	98.52	90.95	49.99	98.9	98.29	99.37	75.89	95.64	95.87	96.6	90