18. Logistic regression and generalized linear models
Model counts, probabilities, and other non-Gaussian outcomes with logistic, Poisson, and related generalized linear models. Produce calibrated class probabilities and explain how features affect predictions.