Table 3.

The results of performance of machine learning classifiers using a combination of multiple questionnaires

	SVM-RBF	SVM-Lin	SVM-Poly	LDA	RF	kNN-2	kNN-5	kNN-10	LR
BDI + HADS
Accuracy	0.715	0.692	0.671	0.678	0.687	0.624	0.670	0.684	0.737
AUC	0.793	0.768	0.745	0.737	0.749	0.658	0.719	0.761	0.812
Sensitivity	0.709	0.687	0.653	0.674	0.686	0.600	0.637	0.650	0.756
Specificity	0.730	0.707	0.714	0.689	0.704	0.708	0.761	0.775	0.723
BDI + PHQ-9
Accuracy	0.695	0.687	0.656	0.677	0.655	0.607	0.650	0.664	0.715
AUC	0.762	0.767	0.719	0.749	0.698	0.635	0.680	0.715	0.782
Sensitivity	0.691	0.677	0.640	0.667	0.658	0.589	0.623	0.634	0.730
Specificity	0.708	0.709	0.691	0.697	0.668	0.665	0.723	0.745	0.702
HADS + PHQ-9
Accuracy	0.692	0.701	0.673	0.701	0.687	0.602	0.650	0.669	0.719
AUC	0.772	0.781	0.744	0.779	0.745	0.639	0.704	0.736	0.794
Sensitivity	0.688	0.703	0.657	0.701	0.675	0.582	0.620	0.636	0.728
Specificity	0.704	0.705	0.709	0.707	0.726	0.678	0.740	0.765	0.714
BDI + HADS + PHQ-9
Accuracy	0.710	0.694	0.674	0.673	0.687	0.607	0.647	0.654	0.734
AUC	0.787	0.774	0.745	0.731	0.743	0.636	0.694	0.730	0.807
Sensitivity	0.706	0.689	0.655	0.667	0.679	0.586	0.618	0.625	0.749
Specificity	0.722	0.708	0.715	0.687	0.715	0.685	0.737	0.737	0.723

BDI, Beck Depression Inventory; HADS, Hospital Anxiety Depression Scale; PHQ-9, The Patient Health Questionnaire-9; AUC, area under the curve; SVM-RBF, support vector machine-radial basis function; SVM-Lin, support vector machine- linear kernel; SVM-Poly, support vector machine-polynomial kernel; LDA, Linear Discriminant Analysis; RF, Random Forest; kNN-n, k-Nearest Neighborhood with k value of n; LR, Logistic Regression.