I have a a simple function that generates a time series square wave:
genseq <- function()
{
x <- numeric(4*365)
s <- seq(as.Date("2005-01-01"), as.Date("2008-12-31"), by="month")
ob <- as.vector(s[c(10,22,34,46)] - as.Date("2005-01-01"))
oe <- as.vector(s[c(11,23,35,47)] - as.Date("2005-01-01"))
for(.index in 1:length(ob))
{
x[ob[.index]:(oe[.index] - 1)] <- 1
}
return(ts(x, frequency=365))
}
Then I take the invenrse:
fs <- fft(series)
fi <- fft(fs, inverse=TRUE) / length(fs)
plot(fi)
This as expected gives me back exactly what I put in. The x axis is label in
Time and the y axis is labeled as 'fi'. Now if I try some filtering and select
a few frequencies:
ff <- complex(length(fs))
sei <- order(e)
ff[sei[1:40]] <- fs[sei[1:40]]
fi <- fft(ff, inverse=TRUE) / length(ff)
plot(fi, main = "After filter")
Now I ge the x-asis as the real component and y-axis as imaginary component.
When does the display switch? How many frequencies do I select before I get the
'Time' axis back? From the above I am selecting 40 (out of 1500).
Thank you.
Kevin
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
