Skip to main content
Fig. 5 | BMC Genomics

Fig. 5

From: Integration of proteomic and metabolomic characterization in atrial fibrillation-induced heart failure

Fig. 5

Biomarkers screening based on machine learning classifier. A Meandecrease accuracy of 10 candidate biomarkers in random forest analysis. B Mean decrease gini of 10 candidate biomarkers in random forest analysis. C Predictive performances of the 10 biomarkers. D ROC curve for prediction with the combined biomarkers. E ROC curve for prediction with NT-proBNP. F Relative concentrations of 10 biomarkers selected by random forest and concentration of NT-proBNP

Back to article page