# R/Gelfand.Diagnostic.R In LaplacesDemon: Complete Environment for Bayesian Inference

#### Documented in Gelfand.Diagnostic

```###########################################################################
# Gelfand.Diagnostic                                                      #
#                                                                         #
# The Gelfand.Diagnostic function is an interpretation of Gelfand's       #
# ``thick felt-tip pen'' MCMC convergence diagnostic (Gelfand et al.,     #
# 1990).                                                                  #
###########################################################################

Gelfand.Diagnostic <- function(x, k=3, pen=FALSE)
{
### Initial Checks
if(missing(x)) stop("The x argument is required.")
if(!is.vector(x)) x <- as.vector(x)
if(k < 2) k <- 2
if(k > length(x)/2) k <- round(length(x)/2)
if({length(x)/k} < 2) stop("k is too large relative to length(x).")
### KDE
quantiles <- seq(from=0, to=1, by=1/k)
breaks <- round(as.vector(quantiles)*length(x))
breaks <- breaks[-1]
d.temp <- density(x)
d <- array(c(d.temp\$x, d.temp\$y), dim=c(length(d.temp\$x), 2,
length(breaks)))
d.temp <- density(x[1:breaks[1]])
d[,,1] <- c(d.temp\$x, d.temp\$y)
for (i in 2:length(breaks)) {
d.temp <- density(x[1:breaks[i]])
d[,,i] <- c(d.temp\$x, d.temp\$y)}
### Plots
ymax <- max(d[,2,])
col.list <- c("red", "green", "blue", "yellow", "purple", "orange",
"brown", "gray", "burlywood", "aquamarine")
col.list <- rep(col.list, len=length(breaks))
rgb.temp <- as.vector(col2rgb(col.list[1]))
mycol <- rgb(red=rgb.temp[1], green=rgb.temp[2], blue=rgb.temp[3],
alpha=50, maxColorValue=255)
plot(d[,1,1], d[,2,1], type="l", col=mycol, xlim=c(range(d[,1,])),
ylim=c(0,ymax), main="Gelfand Diagnostic",
xlab=deparse(substitute(x)), ylab="Density")
polygon(x=d[,1,1], y=d[,2,1], col=mycol, border=NULL)
for (i in 2:length(breaks)) {
rgb.temp <- as.vector(col2rgb(col.list[i]))
mycol <- rgb(red=rgb.temp[1], green=rgb.temp[2],
blue=rgb.temp[3], alpha=50, maxColorValue=255)
lines(d[,1,i], d[,2,i], col=mycol)
polygon(x=d[,1,i], y=d[,2,i], col=mycol, border=mycol)
lines(d[,1,i], d[,2,i], lty=i)}
if(pen == TRUE) abline(v=mean(range(d[,1,])), col="black", lwd=10)
legend(quantile(d[,1,], probs=0.025), round(ymax*0.9,2),
legend=paste("1:",breaks,sep=""), lty=1:k, title="Samples")
return(invisible(x))
}

#End
```

## Try the LaplacesDemon package in your browser

Any scripts or data that you put into this service are public.

LaplacesDemon documentation built on July 1, 2018, 9:02 a.m.