Decision boundaries, cross-entropy, thresholds, and the metrics that matter when classes are imbalanced.
10h · 8 lessons · 1 challenge
Before this: Probability Foundations, Regression & the Supervised Setup
Accuracy is the wrong metric and you will be asked why. This module covers logistic regression as the canonical classifier, then spends its second half on evaluation: precision, recall, ROC, PR curves, calibration and threshold selection under a real cost asymmetry.
Class imbalance gets its own treatment, because almost every interesting production problem has it.
A fraud model, a false-positive cost and a false-negative cost. Move the threshold; the game bills you for every mistake.
Worth 400 XP.
