Exploring Built Environment Influence on Taxi Vacant Time in Megacities: A Case Study of Chongqing, China

Table 2

Estimation results of GAM model.

Explanatory variables	Coefficients
	Estimate	Std. Error	t value	Pr (>|t|)
Intercept	2.0860	0.0083	251.450	<2e − 16
ENT	−0.0027	0.0002	−10.943	<2e − 16
FAI	0.0027	0.0007	3.751	0.0002
GOV	−0.0013	0.0003	−4.339	1.43e − 05
HTS	0.0043	0.0004	10.734	<2e − 16
LFS	0.0018	0.0003	6.787	1.14e − 11
HCS	0.0034	0.0004	8.486	<2e − 16
MLB	−0.0012	0.0012	−0.955	0.3395
SAE	0.0001	0.0003	0.400	0.6893
SHP	−0.0002	0.0001	−3.761	0.0002
SLE	0.0099	0.0005	20.643	<2e − 16
SNS	−0.0007	0.0007	−1.019	0.3081
PMT	−0.0365	0.0081	−4.487	7.23e − 06
MET	0.3239	0.0075	42.915	<2e − 16
Explanatory variables	Approximate significance of smoothing terms
	EDF	Ref. Df	F	value
s (TFS)	8.929	8.999	5393.5	<2e − 16
s (DSR)	8.927	8.998	1711.3	<2e − 16
s (BSN)	8.774	8.978	391.6	<2e − 16
s (CAT)	8.943	8.999	421.6	<2e − 16
s (MVS)	8.949	8.999	687.1	<2e − 16
s (RES)	8.591	8.931	405.5	<2e − 16
s (longitude, latitude)	28.945	29.000	1835.5	<2e − 16
t2 (DOW, TSLOT)	21.966	23.236	117.9	<2e − 16
Model fit	R-sq.(adj)	0.759

Significance codes: “” 0.001. s ( ): smooth spline function. t2 ( ): smooth tensor product.