Skip to content

Lomb periodograms

2 messages · Garcia Carreras, Bernardo, Josué Polanco

#
Hi,

I have recently used the CTS package in order to use the Lomb-Scargle periodogram (spec.ls) function. I have noticed an issue that I hoped you may be able to explain. If a regularly spaced time series has two points removed, one at either side of a single data point (thus making an irregularly spaced time series), a spectrum with a very large peak at the highest frequencies is produced. An example of this is shown below:
	
a <- runif(100)
x <- 1:length(a)

a <- a[-c(4,6)]
x <- x[-c(4,6)]

spec.ls(x=x, y=a, fast=FALSE, taper=0)

Does anyone know why this is? Thank you very much in advance for any help you can give me with this issue.

Best regards,

Bernardo
#
Hi All, hi Bernardo,

I think that the problem is when you remove 2*n data, then the problem
is when the number of the frequencies is computed. Look in the source
code, may be there is the key. If you have further  problems, you
could consider to use a Lomb-Scargle function that I wrote, send me an
email if you want.  Remember that the RAW Lomb-Scargle periodogram is
not a good  estimation  of the spectral power.

Cheers, and good luck,

--
Josue Polanco





On Tue, Mar 31, 2009 at 6:33 PM, Garcia Carreras, Bernardo
<bernardo.garcia-carreras08 at imperial.ac.uk> wrote: