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

Table 1

Variables of road accidents, Adelaide CBD, 2006–2008.

Variable class	Variable name	Discretization value	Value description	Frequency	Percentage
Driver	Apparent error	1	Fail to stand	307	13.25%
		2	Change lanes to endanger	221	9.54%
		3	Incorrect turn	31	1.34%
		4	Reverse without due care	92	3.97%
		5	Follow too closely	173	7.47%
		6	Overtake without due care	52	2.24%
		7	Disobey traffic lights	171	7.38%
		8	Disobey stop sign	23	0.99%
		9	Disobey give way sign	47	2.03%
		10	Inattention	923	39.84%
		11	DUI	23	0.99%
		12	Fail to give way	254	10.96%
Road	Road geometry	1	Cross road	1116	48.17%
		2	Y junction	57	2.46%
		3	T junction	450	19.42%
		4	Multiple	33	1.42%
		5	Divided road	349	15.06%
		6	Not divided	294	12.69%
		7	Pedestrian crossing	18	0.78%
	Road moisture condition	1	Wet	237	10.23%
		2	Dry	2080	89.77%
	Traffic control	1	Traffic signals	1282	55.33%
		2	Stop sign	42	1.81%
		3	Give way sign	118	5.09%
		4	No control	875	37.76%
Environment	Weather condition	1	Raining	151	6.52%
		2	Not raining	2166	93.48%
	Light condition	1	Daylight	1716	74.06%
		2	Night	601	25.94%
Vehicle	Vehicle type	1	Heavy	172	7.42%
		2	Medium	901	38.89%
		3	Light	1244	53.69%
	Vehicle movement	1	Right turn	426	18.39%
		2	Left turn	95	4.10%
		3	U turn	122	5.27%
		4	Swerving	242	10.44%
		5	Reversing	77	3.32%
		6	Straight ahead	1238	53.43%
		7	Entering private driveway	15	0.65%
		8	Leaving private driveway	50	2.16%
		9	Overtaking on right	39	1.68%
		10	Overtaking on left	13	0.56%
Road crash	Crash type	1	Rear end	1010	43.59%
		2	Hit fixed object	59	2.55%
		3	Side swipe	392	16.92%
		4	Right angle	377	16.27%
		5	Head on	5	0.22%
		6	Hit pedestrian	50	2.16%
		7	Right turn	333	14.37%
		8	Hit parked vehicle	91	3.93%
	Crash severity	1	PDO (property damage only)	1748	75.44%
		2	Injury	569	24.56%
	Total units (involved in a road crash)	1	Two units	2012	86.84%
		2	Three units	258	11.14%
		3	Four units	40	1.73%
		4	Five units	7	0.30%
	Total casualties (fatalities and treated injuries)	1	None	1748	75.44%
		2	One casualty	495	21.36%
		3	Two casualties	62	2.68%
		4	Three casualties	12	0.52%
	Total serious injuries	1	None	2267	97.84%
		2	One serious injury	50	2.16%
	Total estimated damage (A$)	1	[0, 5000)	1139	49.16%
		2	[5000, 10000)	795	34.31%
		3	[10000, +∞)	383	16.53%