imf.test: IMF significance test

In PierreMasselot/emdr: Tools for EMD-regression

Description

Tests the hypothesis that IMFs do not contain more information than white noise.

Usage

imf.test(
  object,
  tt = NULL,
  type = c("wu", "flandrin"),
  alternative = c("one.sided", "two.sided"),
  H = 0.5
)

## S3 method for class 'imftest'
print(x, digits = 3, ...)

Arguments

object	mimf object or array containing (potentially multivariate) IMFs.
tt	Numeric vector of custom time indices.
type	The type of significance test to perform. Either the test of Wu and Huang (2004) (type = "wu") or the test of Flandrin et al. (2004) (type = "flandrin").
alternative	Character giving if the alternative hypothesis is one-sided (alternative = "one.sided") or two-sided (alternative = "two.sided").
H	Numeric value giving the Hurst exponent necessary for the test of Flandrin et al. (2004). Must be between 0 and 1.
x	An imftest object resulting from a call to imf.test.
digits	The number of digits to print p-values.
...	Formatting arguments to be passed to the function format

Details

In both tests, the first (highest frequency) IMF is used to estimate the noise level of the input signal and is thus never significant.

Value

An object of class imftest containing the following elements:

object	The input mimf object name.
alternative	The alternative hypothesis.
type	The type of test performed.
logp	Matrix containing the log mean-period of each IMF.
loge	Matrix containing the log energy (variance) of each IMF.
p.value	Matrix containing the p-value of the test for each IMF.

References

Wu, Z., Huang, N.E., 2004. A study of the characteristics of white noise using the empirical mode decomposition method. Proceedings of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences 460, 1597-1611.

Flandrin, P., Goncalves, P., Rilling, G.G., 2004. Detrending and denoising with empirical mode decompositions. Presented at the 12th European Signal Processing Conference, IEEE, Vienna, Austria.

Examples

   library(dlnm)
   
   # Decompose both temperature and relative humidity with NA-MEMD
   # Adding two noise variables 
   X <- chicagoNMMAPS[,c("temp", "rhum")]
   set.seed(123)
   mimfs <- memd(X, l = 2) # Takes a couple of minutes
   cmimfs <- combine.mimf(mimfs, list(10:11, 12:13), 
     new.names = c("C10", "C11"))
   
   # Apply the test of Wu and Huang (2004)
   testres <- imf.test(cmimfs)
   testres
   plot(testres)

PierreMasselot/emdr documentation built on June 19, 2021, 2:11 p.m.