Advertisements

In tomorrow’s class, I’ll hand out the graph below that shows a Bayesian “triplot” (showing the likelihood, prior, and posterior) and the prior predictive distribution. The triplot is useful to show how the the two types of information (data and prior) are combined. The predictive plot is helpful in judging the suitability of the Bayesian model.

I wrote two short R functions, triplot and predplot, that construct the graphs. I assume that you have already loaded the LearnBayes package.

> library(LearnBayes)

> source(url(“http://bayes.bgsu.edu/m648/triplot.R”))

> source(url(“http://bayes.bgsu.edu/m648/predplot.R”))

> prior=c(6.8,2.5)

> data=c(9,15)

> n=sum(data)

> par(mfrow=c(2,1))

> triplot(prior,data)

> predplot(prior,n,data[1])

Advertisements