An R package for Bayesian logistic regression. The posterior distribution of the parameters is obtained via Gibbs sampling using Polya-Gamma latent variables (see paper "Bayesian Inference for Logistic Models Using Pólya–Gamma Latent Variables" for details).
This package can be installed as follows:
devtools::install_github("kasparmartens/PolyaGamma")A small example, how to use the main function gibbs_sampler
# load library
library(PolyaGamma)
# generate data from the logistic model 1 / (1 + exp(-(1 + x)))
data = generate_from_simple_logistic_model(n=100)
# obtain a sample from the posterior distribution of beta
obj = gibbs_sampler(data$y, data$X, lambda=0.001, n_iter_total=200, burn_in=50)
obj
plot(obj)