Code to reproduce simulation results presented in
Centered Partition Processes: Informative Priors for Clustering (with Discussion)
by Sally Paganin, Amy H. Herring, Andrew F. Olshan, David B. Dunson
Bayesian Analysis. 16(1), 301-370, (March 2021) [ Link | arXiV preprint ]

Be sure to have installed the following R packages

	install.packages("Rcpp")
	install.packages("RcppArmadillo")
	install.packages("devtools")
	install.packages("ggplot2")
	install.packages("reshape2")
	install.packages("magrittr")
	install.packages("dplyr")
	install.packages("sdols")
	install.packages("mcclust")
	install.packages("mvtnorm")

Download and install the package CPLogit using devtools

library(devtools)
devtools::install_github("salleuska/CPLogit", subdir="CPLogit")

or if you download the repo

## How to install 
library(devtools)
library(Rcpp)
clean_dll()

## to export Rcpp functions
compileAttributes()
build()
install()
document()

To reproduce code of the prior probability distribution over the set partition space induced by the CP process, use script Sec3.3_Prior_graphs.R.

Simulate data used in the Sec 5.2 and in Sec 1 of the Supplementary Materials

Rscript simulation/0_generateData_linearSimulation.R
Rscript simulation/0_generateData_logisticSimulation.R

Run models and reproduce plots in Sec 5.2 and in Sec 1 of the Supplementary Materials

Rscript simulation/1_linearSimulation.R
Rscript simulation/1_logisticSimulation.R