Towards Driver's State Recognition on Real Driving Conditions

Table 5

Results for the stress classification problem using three feature sets. For each classifier, the sensitivity (Sens.) and the specificity (Spec.) per class are given as well as the total accuracy (Acc.).

		Normal		Stress
		Sens.	Spec.	Sens.	Spec.	Acc.
Physiological features	SVM	0.79	0.78	0.78	0.79	0.78
	Decision Trees	0.78	0.76	0.75	0.77	0.76
	Naive Bayes	0.79	0.63	0.54	0.72	0.66
	Bayes Classifier	0.85	0.63	0.49	0.77	0.67
		Sens.	Spec.	Sens.	Spec.	Acc.
Physiological + video features	SVM	0.80	0.90	0.91	0.82	0.86
	Decision Trees	0.80	0.81	0.81	0.80	0.81
	Naive Bayes	0.82	0.71	0.66	0.79	0.74
	Bayes Classifier	0.85	0.71	0.65	0.82	0.75
		Sens.	Spec.	Sens.	Spec.	Acc.
Physiological + video + environmental features	SVM	0.88	0.85	0.84	0.88	0.86
	Decision Trees	0.82	0.81	0.80	0.81	0.81
	Naive Bayes	0.85	0.76	0.74	0.83	0.79
	Bayes Classifier	0.86	0.76	0.73	0.83	0.79