I was having trouble finding a way to make a binary prediction because my target feature was skewed 97:3, so I made this to compare and choose the best model. It is especially helpful as I work more with the data and need quick visuals.
Heatmap: Overall comparision between models that displays each metric.
- Features a 'hover' function that changes the graphs that are displayed below.
ROC Curve: Plots the true positive rate against the false positive rate
- Includes the AUC, which indicates how well the model is at separating classes
Precision-Recall Curve: Plots precision vs recall.
- Good for when you want to optimize one or the other
Classification Report: Displays precision, recall, f1, and the binary support.
Confusion Matrix: Displays the TP, FP, TN, and FN values for the given model.