pmlr: Penalized Multinomial Logistic Regression

Extends the approach proposed by Firth (1993) for bias reduction of MLEs in exponential family models to the multinomial logistic regression model with general covariate types. Modification of the logistic regression score function to remove first-order bias is equivalent to penalizing the likelihood by the Jeffreys prior, and yields penalized maximum likelihood estimates (PLEs) that always exist. Hypothesis testing is conducted via likelihood ratio statistics. Profile confidence intervals (CI) are constructed for the PLEs.

Install the latest version of this package by entering the following in R:
AuthorSarah Colby <>, Sophia Lee, Juan Pablo Lewinger, Shelley Bull <>
Date of publication2010-04-02 17:08:39
MaintainerSarah Colby <>
LicenseGPL (>= 2)

View on CRAN

Questions? Problems? Suggestions? or email at

Please suggest features or report bugs with the GitHub issue tracker.

All documentation is copyright its authors; we didn't write any of that.