Learn R Programming

mcmcse

An R package for computing Monte Carlo standard errors (MCSE) in Markov chain Monte Carlo (MCMC) settings. MCSE computation for expectation and quantile estimators is supported as well as multivariate estimations. The package also provides functions for computing effective sample size and for plotting Monte Carlo estimates versus sample size.

Installation

This R package is on CRAN, and its preferred URL being https://CRAN.R-project.org/package=mcmcse.

To download this development repo, through the the devtools package:

# install.packages("devtools")
library(devtools)
devtools::install_github("dvats/mcmcse")

Citation

Please run citation("mcmcse") after loading the package for citation details.

Copy Link

Version

Install

install.packages('mcmcse')

Monthly Downloads

1,832

Version

1.5-0

License

GPL (>= 2)

Maintainer

Dootika Vats

Last Published

September 9th, 2021

Functions in mcmcse (1.5-0)

Effective Sample Size of a multivariate Markov chain as described in Vats et al. (2015).

Apply mcse.q to each column of a matrix or data frame of MCMC samples.

Minimum effective sample size required for stable estimation as described in Vats et al. (2015)

Apply mcse to each column of the MCMC samples.

Multivariate Monte Carlo standard errors for expectations.

Compute Monte Carlo standard errors for quantiles.

Confidence regions (ellipses) for Monte Carlo estimates

MCMC samples from a bivariate normal distribution

Check if the class of the object is mcmcse

Compute Monte Carlo standard errors for expectations.

Multivariate Monte Carlo standard errors for expectations with the initial sequence method of Dai and Jones (2017)

Monte Carlo Standard Errors for MCMC

Batch size (truncation point) selection

Univariate effective sample size (ESS) as described in Gong and Flgal (2015).

Create a plot that shows how Monte Carlo estimates change with increasing sample size.