visualizeBF: Function for exploring Bayes Factors.

In mcbeem/visualizeBF: Function for Visualizing the Bayes Factor for the Mean of Continuous Data

Description

Function for exploring Bayes Factors.

Usage

visualizeBF(data, pointsize = 0.001, scale = 0.707, plot = 1)

Arguments

data	A vector of data
pointsize	Width of points for numerical integration, defaults to .001.
scale	Scale parameter for Cauchy prior, defaults to .707
plot	The type of plot to produce. 1=BF as weighted likelihood, 2=BF as comparison of densities under prior and posterior at mu=0. for mu=0 under prior and posterior distributions.

Value

Returns a list with the following components:

• L.m0 The likelihood evaluated at mu=0; the L of the null model (m0)
  

• L.m1 The likelihood integrated over the whole space weighted by m1 prior; likelihood of the alternative model (m1).
  

• prior.mu0 The density of the prior distribution evaluated at mu=0 (m1).
  

• posterior.m0 The density the posterior distribution evaluated at mu=0 (m0).
  

• BF10 The Bayes Factor for alternative hypothesis relative to null.
  

• BF01 The Bayes Factor for the null hypothesis relative to the alternative.
  

• figure Object containing the plot.
  

• note Instructions for restoring plot defaults; temporary.
  

Examples

set.seed(1)
data <- rnorm(n=50, mean=.3, sd=1)
visualizeBF(data, plot=1)
visualizeBF(data, plot=2)

a <- visualizeBF(data, plot=1)

# extract density of the prior distribution at mu=0
a$prior.mu0

# calculate BF10 as likelihood ratio
a$L.m1 / a$L.m0

# alternative method for calculating BF10 as density ratio
a$prior.mu0 / a$posterior.mu0

# view the plot
a$figure

# restore plot defaults
dev.off()

mcbeem/visualizeBF documentation built on Nov. 19, 2019, 12:35 a.m.