Day 1 - Classification Evaluation Metrics: Accuracy, Precision, Recall, F1 and Confusion Matrix

Models trained and evaluated — but "is this model good enough?" requires quantifying. For classification, accuracy isn't one number. Especially in imbalanced problems (fraud 1%, disease diagnosis) accuracy is a trap. Today covers the confusion matrix root, then precision, recall, F1, when to use each based on business cost.

Confusion Matrix: All Classification Metrics Stem Here

Binary classification's every metric derives from this 2×2 table.

                Predicted: Positive    Predicted: Negative
Actual: Positive    TP (True Positive)   FN (False Negative)
Actual: Negative    FP (False Positive)  TN (True Negative)