The multiridge package efficiently estimates multiple penalty parameters (e.g. for different types of covariates) for linear, logistic, and Cox ridge regression. Either cross-validation or empirical Bayes is used for the estimation.