Table 2 Confusion matrix analysis for all datasets (SVM classifier).

Dataset	Method	TP (Collision)	TN (Non-Collision)	FP (Non-Collision)	FN (Collision)	Sensitivity (Recall)	Specificity (Precision)
NASS GES Crash	Proposed	920	8500	150	80	0.92	0.98
	GAN	850	8400	250	150	0.85	0.97
	SMOTE	800	8300	350	200	0.8	0.96
	ENN	750	8200	400	250	0.75	0.95
KSFDDD	Proposed	880	8200	200	100	0.9	0.98
	GAN	800	8100	300	180	0.82	0.96
	SMOTE	750	8000	400	250	0.75	0.95
	ENN	700	7900	500	300	0.7	0.94
Swiss Road Traffic Accident	Proposed	780	7600	150	70	0.92	0.98
	GAN	700	7500	250	150	0.82	0.97
	SMOTE	650	7400	350	200	0.76	0.95
	ENN	600	7300	400	250	0.71	0.94
UK Road Safety	Proposed	850	8300	200	100	0.89	0.98
	GAN	780	8200	300	180	0.81	0.96
	SMOTE	730	8100	400	250	0.75	0.95
	ENN	680	8000	500	300	0.69	0.94

Quick links

Search