Method of Depression Classification Based on Behavioral and Physiological Signals of Eye Movement

Table 4

Comparison of depression classification methods based on fusion signals (behavior and physiology) and different classification models.

	Accuracy (%)	Sensitivity (%)	Specificity (%)	F1 (%)
SVM	88.95	90.48	88.23	87.86
KNN	90.21	90.74	90.60	89.56
Random forests	91.37	91.32	90.96	90.78
KELM	91.00	91.06	92.31	91.29