Table 1.

Model performance indicators of the supervised model for prediction of stress.

Scenarios			Random forest		Support vector machine		Naive Bayes		Logistic regression
Mental issues less than 1 day in a week (class 1) vs mental issues more than 5 days in a week (class 0)
	Accuracy (±CI)	0.80 (0.016)		0.80 (0.016)		0.77 (0.017)		0.77 (0.017)
	Sensitivity (±CI)	0.59 (0.063)		0.56 (0.063)		0.59 (0.063)		0.59 (0.063)
	Specificity (±CI)	0.82 (0.016)		0.82 (0.016)		0.79 (0.017)		0.78 (0.017)
	AUROCa (±CI)	0.71 (0.026)		0.69 (0.026)		0.69 (0.025)		0.68 (0.025)
Mental issues less than 1 day in a week (class 1) vs mental issues more than 3 days in a week (class 0)
	Accuracy (±CI)	0.72 (0.018)		0.72 (0.018)		0.74 (0.017)		0.73 (0.018)
	Sensitivity (±CI)	0.6 (0.041)		0.6 (0.041)		0.56 (0.041)		0.57 (0.041)
	Specificity (±CI)	0.75 (0.018)		0.75 (0.018)		0.78 (0.017)		0.76 (0.018)
	AUROC (±CI)	0.68 (0.022)		0.67 (0.022)		0.67 (0.022)		0.67 (0.022)
Mental issues less than 1 day in a week (class 1) vs mental issues more than 1 day in a week (class 0)
	Accuracy (±CI)	0.66 (0.019)		0.66 (0.019)		0.65 (0.019)		0.62 (0.019)
	Sensitivity (±CI)	0.48 (0.027)		0.49 (0.027)		0.45 (0.026)		0.61 (0.026)
	Specificity (±CI)	0.77 (0.018)		0.76 (0.018)		0.77 (0.018)		0.64 (0.020)
	AUROC (±CI)	0.62 (0.019)		0.62 (0.019)		0.61 (0.020)		0.62 (0.018)

^aAUROC: area under the receiver operating characteristic curve.