Regression

How regression is formalized, why ordinary least squares is maximum likelihood under Gaussian noise, and how ridge, lasso, basis expansions, kernels, and Gaussian processes fit into one picture.