Dan Weitzenfeld <[email protected]>:
How about not dropping data and doing a discrete Fourier transform?
How many obs do you have? If you have T obs, you can get T/2 series
with different frequencies, and you should see an amplitude spike for
the one with periodicity at .2-.4 Hz ...or am I misunderstanding your
problem? What is your analysis intended to achieve, and how?
http://en.wikipedia.org/wiki/Discrete_Fourier_transform
http://www.jstor.org/stable/2660649?seq=17
On Tue, Jun 17, 2008 at 6:39 PM, Dan Weitzenfeld
<[email protected]> wrote:
> Hi Folks,
> I'm working with 60 hz data, and I'm checking for periodicity at the
> .2-.4 Hz range. When I use the command -pergram-, the frequencies I
> would like to look at are scrunched up in the far left side of the
> graph, (I think) because the frequency of my data so dominates the
> frequency of interest.
> My knowledge of signal processing is very limited. Is there an easy
> way to get confirm my hypothesis of periodicity at .2-.4 Hz given my
> situation?
> The workaround I am using is cutting down the density of my data to 30
> Hz, then 15 Hz, then 7.5 Hz by simply dropping every other observation
> each time, and redefining my -tsset- variable accordingly. I fear I
> may be opening myself up to aliasing effects by doing this.
> Any feedback would be much appreciated, even if it is just a pointer
> to a good web resource.
> Best,
> Dan
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
