An Air Traffic Controller Action Extraction-Prediction Model Using Machine Learning Approach

Table 6

Feature importance of random forest models for action value prediction.

Model RR1	FI_RR1 (%)	Model RR2	FI_RR2 (%)	Model RR3	FI_RR3 (%)
Speed	34.2	D_WSSS	82.7	Course	46.6
D_WSSS	32.1	V_Speed	5.2	Longitude	30.1
Altitude	6.3	Altitude	4.0	Latitude	10.4
T_Remain	5.0	T_Remain	2.7	D_RPLL	1.7
V_Speed	3.2	Speed	0.8	T_Remain	1.6
Longitude	2.9	Course	0.8	Altitude	1.4
Course	2.8	Longitude	0.6	D_WSSS	1.1
Latitude	2.1	Latitude	0.5	V_Speed	0.9
Hour	1.7	Hour	0.4	Speed	0.9
Month	1.5	Month	0.3	Hour	0.8