Mapas bivariados usando el paquete classInt

The issue with getting 0s in the confusion matrix results despite having a reasonable AUC might be due to an imbalance in the prediction threshold. By default, the threshold for classifying probabilities into class labels is set to 0.5. If the model is not well-calibrated or if the classes are imbalanced, using 0.5 as the threshold might not be appropriate.

Flexdashboard Interactive Plotting Course

Excluding dels <5 in glucose