sqar: Illustration of some of coda's criterions on the noisy...

In mcsm: Functions for Monte Carlo Methods with R

Description

This function illustrates some of coda's criterions on the noisy squared AR model, using a Metro\-polis-Has\-tings algorithm based on a random walk. Depending on the value of the boolean multies, those criterions are either the geweke.diag and heidel.diag diagnostics, along with a Kolmo\-gorov-Smir\-nov test of our own, or plot(mcmc.list()) if several parallel chains are produced together.

Usage

sqar(T = 10^4, multies = FALSE, outsave = FALSE, npara = 5)

Arguments

T	Number of MCMC iterations
multies	Boolean variable determining whether or not parallel chains are simulated
outsave	Boolean variable determining whether or not the MCMC output is saved
npara	Number of parallel chains

Value

This function produces plots and, if outsave is TRUE, it produces a list with value the MMC sample(s).

Author(s)

Christian P. Robert and George Casella

References

Chapter 8 of EnteR Monte Carlo Statistical Methods

See Also

sqaradap

Examples

ousqar=sqar(outsave=TRUE)

Example output

Loading required package: MASS
Loading required package: coda

Fraction in 1st window = 0.1
Fraction in 2nd window = 0.5 

  var1 
-2.997 

                                   
     Stationarity start     p-value
     test         iteration        
var1 passed       1         0.0666 
                             
     Halfwidth Mean Halfwidth
     test                    
var1 passed    2.25 0.0018   
Type  <Return>   to continue : 

Fraction in 1st window = 0.1
Fraction in 2nd window = 0.5 

  var1 
-1.472 

                                   
     Stationarity start     p-value
     test         iteration        
var1 passed       1         0.423  
                             
     Halfwidth Mean Halfwidth
     test                    
var1 passed    2.25 0.00411  

mcsm documentation built on May 2, 2019, 10:16 a.m.