plotEstimator {BayesSUR}

plot heatmap of estimators

Description

Plot the posterior mean estimators from a BayesSUR class object, including the coefficients beta, latent indicator variable gamma and graph of responses.

Usage

plotEstimator(
  x,
  estimator = NULL,
  colorScale.gamma = grey((100:0)/100),
  colorScale.beta = c("blue", "white", "red"),
  legend.cex.axis = 1,
  name.responses = NA,
  name.predictors = NA,
  xlab = "",
  ylab = "",
  fig.tex = FALSE,
  output = "ParamEstimator",
  header = "",
  header.cex = 2,
  tick = FALSE,
  mgp = c(2.5, 1, 0),
  cex.main = 1.5,
  title.beta = NA,
  title.gamma = NA,
  title.Gy = NA,
  beta.type = "marginal",
  Pmax = 0,
  ...
)

Arguments

x

an object of class BayesSUR

estimator

print the heatmap of estimators. The value "beta" is for the estimated coefficients matrix, "gamma" for the latent indicator matrix and "Gy" for the graph of responses

colorScale.gamma

value palette for gamma

colorScale.beta

a vector of three colors for diverging color schemes

legend.cex.axis

magnification of axis annotation relative to cex

name.responses

a vector of the response names. The default is NA only to show the locations. The value "auto" show the response names from the orginal data.

name.predictors

a vector of the predictor names. The default is NA only to show the locations. The value "auto" show the predictor names from the orginal data.

xlab

a title for the x axis

ylab

a title for the y axis

fig.tex

print the figure through LaTex. Default is FALSE

output

the file name of printed figure

header

the main title

header.cex

size of the main title for all estimators

tick

a logical value specifying whether tickmarks and an axis line should be drawn. Default is FALSE

mgp

the margin line (in mex units) for the axis title, axis labels and axis line

cex.main

size of the title for each estimator

title.beta

a title for the printed "beta"

title.gamma

a title for the printed "gamma"

title.Gy

a title for the printed "Gy"

beta.type

the type of output beta. Default is marginal, giving marginal beta estimation. If beta.type="conditional", it gives beta estimation conditional on gamma=1

Pmax

threshold that truncate the estimator "gamma" or "Gy". Default is 0. If Pmax=0.5 and beta.type="conditional", it gives median probability model betas

...

other arguments

Examples

data("exampleEQTL", package = "BayesSUR")
hyperpar <- list(a_w = 2, b_w = 5)

set.seed(9173)
fit <- BayesSUR(
  Y = exampleEQTL[["blockList"]][[1]],
  X = exampleEQTL[["blockList"]][[2]],
  data = exampleEQTL[["data"]], outFilePath = tempdir(),
  nIter = 10, burnin = 0, nChains = 1, gammaPrior = "hotspot",
  hyperpar = hyperpar, tmpFolder = "tmp/"
)

## check output
# Plot the estimators from the fitted object
plotEstimator(fit, estimator = c("beta", "gamma", "Gy"))

## Not run: 
## Set up temporary work directory for saving a pdf figure
# td <- tempdir()
# oldwd <- getwd()
# setwd(td)

## Produce authentic math formulas in the graph
# plotEstimator(fit, estimator = c("beta", "gamma", "Gy"), fig.tex = TRUE)
# system(paste(getOption("pdfviewer"), "ParamEstimator.pdf"))
# setwd(oldwd)

## End(Not run)