plot3Q: Plot 3-dimensional data modeled with 3 separate structure...

In CARrampsOcl: Reparameterized and marginalized posterior sampling for conditional autoregressive models, OpenCL implementation

Description

Plot 3-dimensional data modeled with 3 separate structure matrices.

Usage

plot3Q(objname, numcols = 64, col = rev(terrain.colors(numcols)), 
plotdims = c(1, 2), rev.inds = c(FALSE, FALSE), blocks = NULL, 
animate = FALSE, intv = 3, 
title = c("Raw data", "Estimated underlying truth"), sub = NULL)

Arguments

objname	name of output object produced by CARrampsOcl.fit
numcols	number of shades from the color palette to be used
col	color palette to be used in plotting; the default plots high values in green and low values in pink.
plotdims	This function produces a sequence of two-dimensional plots. The plotdims argument is a 2-vector identifying which two dimensions appear in the rows and columns of each two-dimensional plot. The numbers refer to the order of the structure matrices specified in the Q argument to CARrampsOcl.fit. The default is the first two dimensions.
rev.inds	Should the plotting indices on each two-dimensional plot be reversed? Setting rev.inds = c(TRUE,FALSE) flips the plots from left to right; rev.inds = c(FALSE,TRUE) turns the plots upside down.
blocks	Subset which two-dimensional plots are displayed. Default is that the corresponding two-dimensional plot is displayed for each value of the third dimension. To subset, set blocks to a set of integers. For example, blocks = 1:5 will display the first 5 plots only.
animate	If animate = FALSE, the user is prompted to press a key to move to the next plot. Otherwise, the next plot will automatically appear after intv seconds.
intv	How many seconds to wait before displaying next plot (ignored if animate=FALSE).
title	Vector of character values; the titles of the two plots.
sub	Vector of subtitles for plots.

Details

For each value of a third dimension, this function plots two two-dimensional plots side-by-side. The left plot is of the raw data input into the CARrampsOcl.fit function, and the right plot is of the estimated means of the of the posterior distributions of the corresponding random effects.

Value

For each value in blocks, this function plots two two-dimensional plots side-by-side. The left plot is of the raw data input into the CARrampsOcl.fit function, and the right plot is of the estimated means of the of the posterior distributions of the corresponding random effects.

Author(s)

Alex Sawyer and Kate Cowles

Examples

## Not run: 
data(iowaSW97_06small)
na<- 17
nb <- 12
nc <- 10
Q <- list( list(type="RW1",content=na), list(type="RW1",content=nb), 
list(type="RW1",content=10))
alpha <- beta <- rep(0.01,4)
X<- matrix( rep(1, na * nb * nc), ncol=1)

# Add noise to data to see whether fitting method can extract true image
y <- iowaSW97_06small/100 + rnorm(na*nb*nc,sd=6)

outputSW <- CARrampsOcl.fit(alpha, beta, Q, y, nsamp=50,                  
     seed=2, fixed = FALSE, coefs = TRUE,
     randeffs = TRUE, designMat = X, mult = 2000, filename = "params.txt")
require(coda)
summary( as.mcmc(1/sqrt(outputSW$params)))

plot3Q( objname=outputSW, numcols=64, col=rev(terrain.colors(64)),
plotdims=c(1,2), rev.inds=c(FALSE,TRUE), blocks=NULL, animate=TRUE,
intv=3, title=c("Raw data","Estimated underlying truth"), sub=1997:2006 )

## End(Not run)

CARrampsOcl documentation built on May 2, 2019, 3:27 a.m.