Replace Boundary Wavelet Coefficients with Missing Values

Sets the first \(n\) wavelet coefficients to NA.

brick.wall(x, wf, method="modwt")
dwpt.brick.wall(x, wf, n.levels, method="modwpt")

Arguments

x	DWT/MODWT/DWPT/MODWPT object
wf	Character string; name of wavelet filter
method	Either dwt or modwt for brick.wall, or either dwpt or modwpt for dwpt.brick.wall
n.levels	depth of the wavelet transform

Value

Same object as x only with some missing values.

Details

The fact that observed time series are finite causes boundary issues. One way to get around this is to simply remove any wavelet coefficient computed involving the boundary. This is done here by replacing boundary wavelet coefficients with NA.

References

Lindsay, R. W., D. B. Percival and D. A. Rothrock (1996). The discrete wavelet transform and the scale anlaysis of the surface properties of sea ice, IEEE Transactions on Geoscience and Remote Sensing, 34, No.~3, 771-787.

Percival, D. B. and A. T. Walden (2000) Wavelet Methods for Time Series Analysis, Cambridge University Press.

Author

B. Whitcher