Here's a reprise on maximum likelihood. maximumLikelihood-reprise.ppt

As background for linear regression, we'll talk about bias versus variance errors and regularization.  Here are the figures from "Elements of Statistical Learning" by Hastie, Tibshirani and Friedman.  figures2.pdf

Here are some rough notes on coefficient shrinkage methods. Glmnet.pdf

This lecture principally covers glmnet.  The basic paper is http://www.jstatsoft.org/v33/i01/ by Professor Jerry Friedman at Stanford.