cvwavelet.after.impute {CVThresh} | R Documentation |
Cross-Validation Wavelet Shrinkage after imputation
Description
This function performs level-dependent cross-validation wavelet shrinkage given the cross-validation scheme and imputation values.
Usage
cvwavelet.after.impute(y, ywd, yimpute,
cv.index, cv.optlevel, cv.tol=0.1^3, cv.maxiter=100,
filter.number=10, family="DaubLeAsymm", thresh.type="soft", ll=3)
Arguments
y |
observation |
ywd |
DWT object |
yimpute |
imputed values according to cross-validation scheme |
cv.index |
test dataset index according to cross-validation scheme |
cv.optlevel |
thresholding levels |
cv.tol |
tolerance for cross-validation |
cv.maxiter |
maximum iteration for cross-validation |
filter.number |
specifies the smoothness of wavelet in the decomposition (argument of WaveThresh) |
family |
specifies the family of wavelets “DaubExPhase" or “DaubLeAsymm" (argument of WaveThresh) |
thresh.type |
specifies the type of thresholding “hard" or “soft" (argument of WaveThresh) |
ll |
specifies the lowest level to be thresholded |
Details
Calculating the threshold values and reconstructing noisy data y
, given the index of each testdata,
imputed values according to cross-validation scheme and discrete wavelet transform of y
.
Value
Reconstruction and thresholding values by level-dependent cross-validation
yc |
reconstruction |
cvthresh |
thresholding values by level-dependent cross-validation |
See Also
cvwavelet
, cvtype
, cvimpute.by.wavelet
.
Examples
data(ipd)
y <- as.numeric(ipd); n <- length(y); nlevel <- log2(n)
set.seed(1)
cv.index <- cvtype(n=n, cv.bsize=2, cv.kfold=4, cv.random=TRUE)$cv.index
yimpute <- cvimpute.by.wavelet(y=y, impute.index=cv.index)$yimpute
ywd <- wd(y)
#out <- cvwavelet.after.impute(y=y, ywd=ywd, yimpute=yimpute,
#cv.index=cv.index, cv.optlevel=c(3:(nlevel-1)))
#ts.plot(ts(out$yc, start=1229.98, deltat=0.02, frequency=50),
# main="Level-dependent Cross Validation", xlab = "Seconds", ylab="")
##### Specifying thresholding structure
# cv.optlevel <- c(3) # Threshold (level 3 to finest level) at the same time.
# cv.optlevel <- c(3, 5) # Threshold two groups of resolution levels,
# (level 3, 4) and (level 5 to finest level).
# cv.optlevel <- c(3,4,5,6,7,8) # Threshold each resolution level 3 to 8.