Wavelet Shrinkage

DESCRIPTION:

Applies nonlinear shrinkage rule to wavelet coefficients. This function is usually used for data smoothing or signal de-noising.

This function is an S Version 3 generic (see Methods); method functions can be written to handle specific S Version 3 classes of data. Classes which already have methods for this function include:
crystal.list, crystal.vector, crystal.matrix, wp.crystal .

USAGE:

shrink(x, threshold, high.threshold=threshold, shrink.fun=NULL) 

REQUIRED ARGUMENTS:

x
any numeric S-PLUS object.
threshold
a non-negative value or a vector, see details below.

OPTIONAL ARGUMENTS:

high.threshold
upper threshold, see details below.
shrink.fun
shrink function: one of the character strings "soft" or "hard". See details below.

VALUE:

an object with the same structure as the input with shrunk values.

DETAILS:

If x is a crystal or a numeric vector, threshold should be a non-negative number. If x is a list, threshold should be a vector. If length(threshold)>=length(x), the last length(x) in threshold vector are used; otherwise, length(x)-length(threshold) zeros are added at the beginning of the threshold vector.

REFERENCES:

Donoho, D. and Johnstone, I. (1992). Ideal Spatial Adaptation by Wavelet Shrinkage. Technical report, Department of Statistics, Stanford University.

SEE ALSO:

, .