A unified approach to regularize maximum likelihood models for linear, multivariate normal, logistic (binomial and polytomous), Poisson, and Cox proportional hazards based on between-predictor correlation is presented. Regularization is accomplished via penalization based on the strength of...