It provides the log-likelihood, gradient and observed information matrix for penalized/unpenalized maximum likelihood optimization when fitting univariate models with discrete/continuous response.
Maintainer: Giampiero Marra giampiero.marra@ucl.ac.uk