dfplotcov: Plot parameter covariance matrix

In obreschkow/dftools: Fitting distribution functions such as galaxy mass functions

Description

This function displays the parameter covariances of the distribution function parameters (e.g. the mass function parameters) fitted using dffit.

Usage

dfplotcov(
  data,
  master = 1,
  order = seq(length(data)),
  names = NULL,
  col = c("blue", "black", "red", "green", "orange"),
  lwd = 1,
  lty = 1,
  cex = 1.2,
  pch = 20,
  cloud.alpha = 0.1,
  cloud.nmax = NULL,
  hist.alpha = 0.4,
  show.histogram = TRUE,
  show.cloud = TRUE,
  show.gaussian = TRUE,
  show.expectation = TRUE,
  show.ellipse.68 = TRUE,
  show.ellipse.95 = TRUE,
  show.xlab = TRUE,
  show.ylab = TRUE,
  title = "",
  nstd = 10,
  pmin = NULL,
  pmax = NULL,
  vert.stretch = 1,
  nbins = NULL,
  lower = TRUE,
  upper = FALSE,
  margins = c(4, 4, 0.5, 0.5),
  text.size.labels = 1.1,
  text.size.numbers = 0.8,
  text.offset.labels = c(0, 0),
  text.offset.numbers = c(0, 0),
  text.format.numbers = "%4.1f"
)

Arguments

data	List of objects to be plotted. Each object in this list must be one of the following four: (1) a vector of parameters; (2) a list of two objects, a vector of parameters and an associated covariance matrix; (3) a matrix, where each row represents a set of model parameters, to be displayed as a point of a cloud; (4) an output list produced by dffit. The plot will be centered on the parameters (or average parameters) of the first object in the data list, unless the argument master is set to a different index.
master	Integer specifying the index of the object in the data list defining the axes scales
order	Vector of integer specifying the order of drawing.
names	Optional list of parameter names
col	Vector of colors of each object in the data-list
lwd	Vector of line width of each object in the data-list
lty	Vector of line types of each object in the data-list
cex	Vector of point sizes for the central parameters each object in the data-list
pch	Vector of point type for the central parameters each object in the data-list
cloud.alpha	Vector of alpha values setting the transparency of the point clouds
cloud.nmax	Vector specifying the maximum number of points to be plotted in a point cloud
hist.alpha	Vector of alpha values setting the transparency of the histograms
show.histogram	= Logical vector specifying whether to plot histograms
show.cloud	Logical vector specifying whether to plot the point cloud
show.gaussian	Logical vector specifying whether to draw the Gaussian apprixmation of the parameter distribution
show.expectation	Logical vector specifying whether to mark the expected values (= averages if a parameter set is specified)
show.ellipse.68	Logical vector specifying whether to draw 68% confidence ellipses
show.ellipse.95	Logical vector specifying whether to draw 95% confidence ellipses
show.xlab	Logical flag indicating whether labels+numbers along x-axis are displayed
show.ylab	Logical flag indicating whether labels+numbers along y-axis are displayed
title	Optional plot title
nstd	Width of the plots in multiples of the standard deviations of the model, i.e. the square roots of the diagonal elements of covariance.
pmin	optional P-vector with lower parameter limits to be potted. If given, pmax must also be given and nstd is ignored.
pmax	optional P-vector with lower parameter limits to be potted. If given, pmin must also be given and nstd is ignored.
vert.stretch	P-vector specifying the vertical stretch of the diagonal panels.
nbins	Optional integer specifying the number of bins used in histograms. If set to NULL, this number is determined automatically.
lower	Logical flag indicating whether the lower triangle is shown
upper	Logical flag indicating whether the lower triangle is shown
margins	Plot margins (bottom,left,top,right)
text.size.labels	Text size of parameter names
text.size.numbers	Text size of numbers
text.offset.labels	2-element vector to adjust the position of the vertical and horizontal parameter names
text.offset.numbers	2-element vector to adjust the position of the vertical and horizontal numbers
text.format.numbers	String specifying the floating point format of the numbers. (see sprintf)

Author(s)

Danail Obreschkow

See Also

See examples in dffit.

Examples

# generate two random correlated vectors p[1,] and p[2,] of n elements
# and display their covariance plot
n = 100
p = array(NA,c(n,2))
p[,1] = rnorm(n)
p[,2] = 0.25*p[,1]+rnorm(n,1,0.5)
dfplotcov(list(p))

# now produce the same plot, but increase the number of bins
# add in red color the theoretical expectation
expected_mean = c(0,1)
expected_covariance = cbind(c(1,0.5^2),c(0.5^2,0.5^2+0.25^2))
dfplotcov(list(p,list(expected_mean,expected_covariance)), nbins=20, col=c('black','red'))

# Fit a Schechter function to a mock survey, plot the best-fitting parameters
# with uncertainties in blue and add input parameters as black crosses
dat = dfmockdata(n=1000,sigma=0.5)
survey = dffit(dat$x, dat$veff, dat$x.err)
ptrue = c(-2,11,-1.3)
dfplotcov(list(survey,ptrue), col=c('blue','black'), pch=c(20,3))

obreschkow/dftools documentation built on June 25, 2021, 10:45 p.m.