pmlr: Penalized Multinomial Logistic Regression

Share:

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.

Author
Sarah Colby <colby@lunenfeld.ca>, Sophia Lee, Juan Pablo Lewinger, Shelley Bull <bull@lunenfeld.ca>
Date of publication
2010-04-02 17:08:39
Maintainer
Sarah Colby <colby@lunenfeld.ca>
License
GPL (>= 2)
Version
1.0

View on CRAN

Man pages

enzymes
Liver Enzyme Data
hepatitis
Post-transfusion hepatitis: impact of non-A, non-B hepatitis...
pmlr
Penalized maximum likelihood estimation for multinomial...

Files in this package

pmlr
pmlr/data
pmlr/data/enzymes.R
pmlr/data/hepatitis.R
pmlr/DESCRIPTION
pmlr/man
pmlr/man/enzymes.Rd
pmlr/man/hepatitis.Rd
pmlr/man/pmlr.Rd
pmlr/NAMESPACE
pmlr/R
pmlr/R/getA.R
pmlr/R/getAstar.R
pmlr/R/getMLEs.R
pmlr/R/getPLEs.R
pmlr/R/getV.R
pmlr/R/getW.R
pmlr/R/misc.R
pmlr/R/pmlr.R
pmlr/R/profile.R
pmlr/R/profileCIs.R
pmlr/R/SASalgo.R
pmlr/R/summary.pmlr.R
pmlr/R/test.LR.R
pmlr/R/test.wald.R