Fig. 2: Receiver operating characteristic and precision-recall curves for prediction of bipolar disorder diagnosis. | Translational Psychiatry

Fig. 2: Receiver operating characteristic and precision-recall curves for prediction of bipolar disorder diagnosis.

From: A machine learning algorithm to differentiate bipolar disorder from major depressive disorder using an online mental health questionnaire and blood biomarker data

Fig. 2: Receiver operating characteristic and precision-recall curves for prediction of bipolar disorder diagnosis.

A, B Out-of-fold results of nested cross-validation in the primary dataset (N = 126 BD previously diagnosed as MDD vs. N = 187 confirmed MDD). Thick lines represent curves calculated from probabilities averaged across all models. C–F Validation in baseline low mood group (N = 98 newly diagnosed BD vs. N = 112 newly diagnosed MDD (C, D) and vs. N = 120 with subclinical depressive symptoms (E, F)). AUROC and AUPRC values represent mean (95% CI). AUPRC area under the precision-recall curve, AUROC area under the receiver operating characteristic curve, BD bipolar disorder, CI confidence intervals, MDD major depressive disorder.

Back to article page