Computational Intelligence and Neuroscience

Review Article

Incident Duration Modeling Using Flexible Parametric Hazard-Based Models

Table 4

Regression coefficients of different factors and the percent change for each incident.


Variable	Preparation time		Travel time		Clearance time		Total time
Variable	Parameter estimation	Percent change (%)	Parameter estimation	Percent change (%)	Parameter estimation	Percent change (%)	Parameter estimation	Percent change (%)

Best model	Generalized gamma		Flexible parametric (df5)		Log-normal		Generalized gamma
Peak hour	—	—	—	—	—	—	—	—
Day first shift	—	—	−0.691 (−7.61)	−49.89	0.553 (5.41)	73.84	0.396 (6.96)	48.58
Weekday	—	—	—	—	—	—	—	—
Summer (reference: spring)	0.125 (5.35)^*	13.31	—	—	—	—	—	—
Autumn (reference: spring)	0.156 (6.15)	16.88	−0.171 (−2.81)		—	—	—	—
Winter (reference: spring)	—	—	—	—	—	—	−0.075 (−2.25)	−7.22
Rear-end (reference: more common collision)	—	—	—	—	—	—	—	—
Bike (people) included (reference: more common collision)	—	—	−0.626 (−2.97)	−46.52	—	—	0.455 (3.46)	57.61
Collision with stationary object (reference: more common collision)	—	—	−0.302 (−2.08)	−26.06	—	—	0.202 (2.20)	22.38
Overturned vehicle (reference: more common collision)	−0.334 (−2.38)	−28.39	—	—	0.967 (2.54)	163.00	—	—
Fire (reference: more common collision)	—	—	—	—	—	—	—	—
Vehicle number	—	—	—	—	—	—	—	—
Taxi	0.045 (2.20)	4.60	0.109 (2.20)	11.51	—	—	—	—
Bus	—	—	0.125 (2.01)	13.31	—	—	—	—
Truck	—	—	—	—	—	—	—	—
Distance	−0.035 (−4.23)	−3.43	—	—	0.255 (10.71)	29.04	0.135 (10.03)	14.45
Congestion	−0.060 (−3.64)	−5.82	—	—	0.174 (3.92)	19.00	0.106 (4.36)	11.18
Preparation time	NA	NA	—	—	−0.017 (−2.05)	−1.68	NA	NA
Travel time	NA	NA	NA	NA	—	—	NA	NA

NA: The factor was not used in the model.
—: The factor was not significant at 95% level of significance.
*: If the table cell included numbers, which means the variables are statistically significant at a 95% confidence level.
(·): The number in () was the statistical magnitude for each parameter estimation.
Other explanation: The numbers in the parameter estimation column indicate the regression coefficients of different factors.
The numbers in the percent change column indicate the factor effect on each incident phase. For the AFT model, the number indicates the percent change in time. For the flexible parametric model, the number indicates the percent change in hazard rate.