sqndwdecomp function

Brute-force calculation of the non-decimated squared wavelet transform.

Brute-force calculation of the non-decimated squared wavelet transform.

Accurate, but brute-force, direct (slow) calculation of the non-decimated squared wavelet transform.

sqndwdecomp(x, J, filter.number, family)

Arguments

  • x: The sequence you want to transform
  • J: The number of resolutions you want
  • filter.number: The wavelet filter you wish to use.
  • family: The wavelet family you wish to use

Details

Works by computing the discrete wavelets and the necessary scales using hwwn.dw function. Then forms the direct inner product with the data with the squares of the discrete wavelets.

Returns

Returns a matrix of J rows and length(x) columns. Row j in the matrix returned corresponds to the level (nlev-j) resolution level coefficients (where nlev is nlevelsWT(x)) in the WaveThresh ordering.

References

Fryzlewicz, P., Nason, G.P. and von Sachs, R. (2008) A wavelet-Fisz approach to spectrum estimation. J. Time Ser. Anal., 29 , 868-880.

Author(s)

Piotr Fryzlewicz (modified by Guy Nason)

See Also

sqndwd

Examples

# # Generate random series and then take transform x <- rnorm(128) y <- sqndwdecomp(x=x, J=2, filter.number=3, family="DaubExPhase")
  • Maintainer: Guy Nason
  • License: GPL-2
  • Last published: 2023-09-13

Useful links