Evaluating Predictive Model Performance Using the Confusion Matrix

The confusion matrix divides the model's output into the four segments listed below.

Key indicators of confusion matrix:

Precision = TP / (TP + FP): The ratio of correct predictions to the total number of predictions.

True Positive Rate (TPR) = TP / (TP + FN): What is the ratio of accurate predictions to all actual positive outcomes?

False Positive Rate (FPR) = FP / (FP + TN): The ratio of false predictions to all actual positive outcomes is known as the False Discovery Rate (FDR), which is calculated by dividing the number of false positives by the sum of false positives and true positives. It is expressed as FDR = FP / (FP + TP).

Alert Rate = (TP + FP) / (TP + FP + TN + FN): The ratio of positive predictions over all predictions.

Accuracy = TP + TN / (TP + TN + FP + FN): What is the ratio of correctly classified transactions to the total number of transactions?

Is it possible to convert key indicators from a confusion matrix into a score? Yes, by using the following method:

F1 Score = 2x (Precision x Recall) / (Precision + Recall=TPR)

Key indicators from a confusion matrix can indeed be transformed into Receiver Operating Characteristics (ROC), which is a plot of the True Positive Rate (TPR) against the False Positive Rate (FPR) at various threshold settings.