Modeling the Frequency of Cyclists’ Red-Light Running Behavior Using Bayesian PG Model and PLN Model

Table 2

Summary of model estimation results.

Variables	Parameters	PG model				PLN model
		Mean	Std.	Percentiles		Mean	Std.	Percentiles
				5%	95%			5%	95%
Interception		−1.470	0.521	−1.513	−1.594	−1.109	0.614	−1.159	−1.204
Bicycle flow		1.676	0.453	1.639	1.811	1.634	0.504	1.593	1.765
Conflict traffic volume		−0.215	0.056	−0.220	−0.233	−0.186	0.069	−0.192	−0.202
Pedestrian signal type		0.514	0.219	0.496	0.555	0.564	0.176	0.550	0.609
Vehicle speed	β4	−0.103	0.058	−0.108	−0.112	−0.112	0.053	−0.116	−0.122
E-bike rate		0.352	0.172	0.338	0.380	0.301	0.337	0.322	0.377
Deviance information criterion	DIC	1097				1063
Dispersion parameter		1.616
Extra-Poisson variation						0.733
Bayesian value		0.481
		0.656
		0.445