readme.md

In mcbeem/plotBayes: Function for exploring Bayesian inference for the sample mean for Normal data

The plotBayes package illustrates Bayesian inference, showing how the prior distribution and the likelihood of the data combine to produce a posterior distribution.

This is mostly a toy package useful for teaching.

Installation

You need the devtools package to install this from github.

install.packages("devtools")

Then install plotBayes.

devtools::install_github("mcbeem/plotBayes")

And then load it.

library(plotBayes)

Example

Normal prior with μ = 0, σ = 0.5:

set.seed(1)
data <- rnorm(n=10, mean=1, sd=1)
plotBayes(data, prior.type="normal", prior.parameters=c(0, .5), min=-2, max=2)

## $data.mean
## [1] 1.132203
## 
## $map
## [1] 0.9109109
## 
## $eap
## [1] 0.9103311
## 
## $credible.interval
## [1] 0.4744745 1.3433433

You can request a different credible interval with with argument credible=.

plotBayes(data, prior.parameters=c(.0, .5), prior.type="normal",
          min=-2, max=2, credible=.68)

## $data.mean
## [1] 1.132203
## 
## $map
## [1] 0.9109109
## 
## $eap
## [1] 0.9103311
## 
## $credible.interval
## [1] 0.6866867 1.1271271

Uniform prior with a = .7, b = 1.5:

plotBayes(data, prior.type="uniform", prior.parameters=c(.7, 1.5), min=-2, max=2)

## $data.mean
## [1] 1.132203
## 
## $map
## [1] 1.131131
## 
## $eap
## [1] 1.120113
## 
## $credible.interval
## [1] -1.951952 -1.239239

mcbeem/plotBayes documentation built on Nov. 19, 2019, 8:15 a.m.