Displaying 1 result from an estimated 1 matches for "parzenwin".
2024 Jul 10
1
Implementation for selecting lag of a lag window spectral estimator using generalized cross validation (using deviance)
...n=length(lh), k=1 , nw=2, returnEigenvalues=TRUE)
# Tapered series
lh * mytaper$v
# I may compute the spectrum with reduced bias as:
plot(spectrum(lh*mytaper$v,method="pgram"),log="dB")
# We now focus on the variance
# For a fixed m = 10, using a Parzen window.
library(gsignal)
parzenwin(10)
# The following 2 lines of code, where I try to do the same thing in 2
different ways, did not work for me:
kernapply( spectrum(lh*mytaper$v,method="pgram")$spec,parzenwin(10),circular=TRUE)
spectrum(lh*mytaper$v,kernel = parzenwin(10),method="pgram")
# ?spec.pgram says...