classical-mlsupervisedregression
Linear regression
OLS, ridge, lasso — the simplest supervised model and the gateway to understanding regularisation.
Уровни глубины
L0Intro~2ч
Fits a line through points; interprets coefficients.
L1Basics~10ч
Derives OLS normal equations, ridge closed form; diagnoses multicollinearity.
L2Working~15ч
Implements lasso via coordinate descent; uses scikit-learn for elastic net with CV.
L3Advanced~25ч
Analyses bias-variance tradeoff analytically; generalised linear models.
L4Research~60ч
High-dimensional statistics, sparse recovery theory.
Ресурсы
L0 — Intro
L2 — Working