plotdenwd

Plot the wavelet coefficients of a p.d.f.

DESCRIPTION

USAGE

REQUIRED ARGUMENTS

wd

Wavelet decomposition object, usually output from denwd, possibly thresholded.

OPTIONAL ARGUMENTS

first.level

This specifies how many of the course levels of coefficients are omitted from the plot. The default value of 0 means that all levels are plotted.

top.level

This tells the plotting rountine the true resolution level of the finest level of coefficients. The default results in the coarsest level being labelled 0. The "correct" value can be determined from the empirical scaling function coefficient object (output from denproj) as in the example below.

main

The title of the plot.

scaling

The type of scaling applied to levels within the plot. This can be "compensated", "by.level" or "global". See plot.wd for further details.

rhlab

Determines whether the scale factors applied to each level before plotting are printed as the right hand axis.

NotPlotVal

If the maximum coefficient in a particular level is smaller than NotPlotVal, then the level is not plotted.

VALUE

Axis labels to the right of the picture (scale factors). These are returned as they are sometimes hard to read on the plot.

DETAILS

Basically the same as plot.wd except that it copes with the zero boundary conditions used in density estimation. Note that for large filter number wavelets the high level coefficients will appear very squashed compared with the low level coefficients. This is a consequence of the zero boundary conditions and the use of the convention that each coefficient is plotted midway between two coefficients at the next highest level, as in plot.wd.

SEE ALSO

EXAMPLES

# Simulate data from the claw density, find the empirical
# scaling function coefficients, decompose them and plot
# the resulting wavelet coefficients
 
 data <- rclaw(100)
 datahr <- denproj(data, J=8, filter.number=2, family="DaubExPhase")
 data.wd <- denwd(datahr)
 plotdenwd(data.wd, top.level=(datahr$res$J-1))



# Now use a smoother wavelet

 datahr <- denproj(data, J=8, filter.number=10, family="DaubLeAsymm")
 data.wd <- denwd(datahr)
 plotdenwd(data.wd, top.level=(datahr$res$J-1))

AUTHOR