Table 3.

Classification performance for each feature set using a logistic regression model^a

Feature Set	Accuracy	Sensitivity	Specificity
All features	0.72	0.75	0.70
GSR	0.78	0.75	0.80
HR	0.66	0.50	0.75
ST	0.59	0.50	0.65
Recovery	0.63	0.50	0.70
Response	0.78	0.75	0.80

Abbreviations: GSR, galvanic skin response; HR, heart rate; ST, skin temperature.

^aThe performance is evaluated using accuracy, sensitivity, and specificity. Classifications based on the GSR and response‐related features give the best performance.