Slides

Lecture slides for each day

The deck used in each lecture, as a PDF. Day 1 sets up the mathematics the course rests on, from vectors and probability to loss and least squares. Day 2 is taught as two lectures: what it means for a machine to learn, how a rule is trained and tested, and what a performance result means; then linear and logistic regression. Day 3 covers trees, forests and boosting. Day 4 covers clustering, PCA and using models with care. Each deck is released as the course reaches that day.