test.data {waveband}

Test functions for wavelet regression and thresholding

Description

This function evaluates the "blocks", "bumps", "heavisine" and "doppler" test functions of Donoho & Johnstone (1994b) and the piecewise polynomial test function of Nason & Silverman (1994). The function also generates data sets consisting of the specified function plus uncorrelated normally distributed errors.

Usage

test.data(type = "ppoly", n = 512, signal = 1, rsnr = 7, plotfn = FALSE)

Arguments

type

Test function to be computed. Available types are "ppoly" (piecewise polynomial), "blocks", "bumps", "heavi" (heavisine), and "doppler".

n

Number of equally spaced data points on which the function is evaluated.

signal

Scaling parameter; the function will be scaled so that the standard deviation of the data points takes this value.

rsnr

Root signal-to-noise ratio. Specifies the ratio of the standard deviation of the function to the standard deviation of the simulated errors.

plotfn

If plotfn=TRUE, then the test function and the simulated data set are plotted.

Value

A list with the following components:

x

The points at which the test function is evaluated.

y

The values taken by the test function.

ynoise

The simulated data set.

type

The type of function generated, identical to the input parameter type .

rsnr

The root signal-to-noise ratio of the simulated data set, identical to the input parameter rsnr .

SIDE EFFECTS

If plotfn=TRUE, the test function and data set are plotted.