penlogit

A Stata user-written command for penalized logistic regression

• Current version: 1.1.0
• Release date: 19jul2017

---

Description of the command

penlogit estimates penalized logistic regression models for a binary response via data augmentation. It allows the user to impose Normal and generalized log-F prior distributions on one or more model parameters (log odds-ratios).

Installation

To install the last version of penlogit from GitHub, type

net install penlogit, from("https://raw.githubusercontent.com/anddis/penlogit/master/")

from within a web-aware Stata.

Authors

Andrea Discacciati and Nicola Orsini (Karolinska Institutet, Stockholm, Sweden)

Sander Greenland (University of California, Los Angeles, CA)

References

Andrea Discacciati, Nicola Orsini, and Sander Greenland. Approximate Bayesian logistic regression via penalized likelihood by data augmentation. The Stata Journal 2015; 15(3):712–736