Table 5.

Comparison of the Performance of the Five Machine Learning Algorithms (Test; n=21)

	Lasso	Ridge	Random Forest	LDA	SVM
% of good classification (95% CI)	85.7% (71.4, 95.2)	81.0% (61.9, 95.2)	76.2% (52.4, 90.5)	81.0% (66.7, 95.2)	85.7% (68.9, 1.00)
Sensitivity	83.3% (60.7, 100)	83.3% (57.7, 100)	83.3% (61.5, 100)	83.3% (60.7, 100)	91.7% (72.0, 100)
Specificity	88.9% (65.1, 100)	77.8% (50.0, 100)	66.7% (26.7, 92.3)	77.8% (46.3, 100)	77.8% (46.3, 100)
ROC curve AUC	0.861 (0.64, 1.00)	0.806 (0.52, 0.98)	0.796 (0.50, 0.94)	0.870 (0.58, 1.00)	0.806 (0.56, 1.00)

Abbreviations: AUC, area under the curve; LDA, linear discriminant analysis; ROC, Receiving Operating Characteristic; SVM, support vector machine.