SMPracticals (version 1.4-3.1)

beaver.gibbs: Gibbs Sampler for Normal Changepoint Model, Practical 11.7

Description

This function implements a Gibbs sampler for the normal changepoint model applied to the beaver temperature data used in Example 6.22 and Practical 11.7 of Davison (2003), which should be consulted for details.

Usage

beaver.gibbs(init, y, R = 10, a = 1, b = 0.05)

Value

A matrix of size R x 6, whose first four columns contain the values of the parameters for the iterations. Columns 5 and 6 contain the log likelihood and log prior for that iteration.

Arguments

init

Initial values for parameters

y

A series of normal observations

R

Number of iterations of sampler

a

Value of a hyperparameter

b

Value of a hyperparameter

Author

Anthony Davison ([e-mail])

Details

This is provided simply so that readers spend less time typing. It is not intended to be robust and general code.

References

Davison, A. C. (2003) Statistical Models. Cambridge University Press. Practical 11.7.

Examples

## From Example 11.7:
data(beaver)
system.time( gibbs.out <- beaver.gibbs(c(36, 40, 3, 38), beaver$temp, R=1000))
par(mfrow=c(2,3))
plot.ts(gibbs.out[,1],main="mu1") # time series plot for mu1
plot.ts(gibbs.out[,2],main="mu2") # time series plot for mu2
plot.ts(gibbs.out[,3],main="lambda") # time series plot for lambda
plot.ts(gibbs.out[,4],main="gamma") # time series plot for gamma
plot.ts(gibbs.out[,5],main="log likelihood")  # and of log likelihood