Predicting peripartum depression using elastic net regression and machine learning: the role of remnant cholesterol

Table 4 Comparison of discrimination characteristics among six machine learning models

Characteristics	RF	AdaBoost	SVM	DT	KNN	LD
Accuracy	0.738	0.569	0.677	0.708	0.677	0.646
Precision /Positive Predictive Value	0.754	0.611	0.683	0.720	0.683	0.653
Recall /True Positive Rate	0.738	0.569	0.677	0.708	0.677	0.646
FPR	0.255	0.409	0.323	0.288	0.323	0.354
F₁ Score	0.738	0.562	0.678	0.710	0.678	0.648
AUC*	0.850	0.655	0.677	0.712	0.677	0.777
NPV	0.749	0.598	0.674	0.706	0.674	0.643
True Negative Rate	0.745	0.591	0.677	0.712	0.677	0.646
False Negative Rate	0.255	0.409	0.323	0.288	0.323	0.354

Abbreviations: EPDS, Edinburgh Postnatal Depression Scale; RF, random forest; AdaBoost, adaptive boosting; SVM, support vector machine; DT, decision tree; KNN, k-nearest neighbors; LD, linear discriminant; FPR, false positive rate; AUC, area under curve; NPV, negative predictive value
Note: *The corresponding p-values and confidence intervals for the above values are not displayed in the JASP platform

ISSN: 1471-2393