A Bayesian Network Approach to Causation Analysis of Road Accidents Using Netica

Table 4

Inference results for variables that are associated with “serious injuries” and “total estimated damage” in serious traffic accidents.

Variable class	Variable name	Serious injuries/%	Total estimated damage (≥10,000 AUD)/%
Driver’s apparent error	Fail to stand	3.54	17.5
	Change lanes to endanger	2.83	17.5
	Incorrect turn	3.61	17.1
	Reverse without due care	4.03	17.9
	Follow too closely	3.91	20.0
	Overtake without due care	3.36	18.7
	Disobey traffic lights	4.15	19.4
	Disobey stop sign	4.81	19.5
	Disobey give way sign	4.68	18.9
	DUI	5.17	20.4
	Fail to give way	3.91	17.2
Road geometry	Cross road	4.47	19.6
	Y junction	3.69	18.2
	T junction	3.27	18.0
	Multiple	4.17	18.4
	Divided road	4.05	18.1
	Not divided	4.12	18.3
Weather condition	Raining	4.22	18.4
	Not raining	3.69	18.1
Light condition	Daylight	3.65	18.1
	Night	3.93	18.2
Crash type	Rear end	3.40	19.0
	Hit fixed object	2.76	17.4
	Side swipe	1.20	16.9
	Right angle	2.48	17.6
	Head on	15.4	26.0
	Hit pedestrian	9.37	18.6
	Right turn	2.93	17.2
	Hit parked vehicle	0.34	16.9