Fig. 6
From: Using random forest and biomarkers for differentiating COVID-19 and Mycoplasma pneumoniae infections
Confusion matrix for the random forest classifier’s performance in predicting patient groups. MP, Mycoplasma pneumoniae; COVID-19 positive group shows 9 correctly classified, 3 misclassified as MP positive, and none as co-infected. MP positive group shows 41 correctly classified, 5 misclassified as COVID-19 positive, and none as co-infected. Co-infected group shows 0 correctly classified, 3 misclassified as COVID-19 positive, and 4 as MP positive. Overall accuracy of the model is 0.77, demonstrating moderate classification accuracy for COVID-19 and MP groups but significant misclassification in co-infected cases.
