comp.lang.idl-pvwave archive: archive » CHISQR

Home » Public Forums » archive » CHISQR_CVF question.

Show: Today's Messages :: Show Polls :: Message Navigator
E-mail to friend

CHISQR_CVF question. [message #67654]

Wed, 19 August 2009 09:12

R.G. Stockwell
Messages: 363
Registered: July 1999

Senior Member

I'm just writing up a simple routine to calculate
significance levels for an FFT of a white spectrum.
I actually find the distribution of the spectrum, normalize
it with respect to N number of points and variation so I
always get the same distribution, and I am comparing it
to the CHISQR_CVF function (with 2 degrees of freedom, for a
power spectrum).

I am off by a constant factor (which depends on degrees of freedom)
but invariant to length of the time series, or the standard deviation
(since those are normalized out).

any ideas of what step i am missing here? I frankly can't think
of any other parameter involved here.

cheers,
bob

PS with degrees of freedom = 2, the constant factor is
10.914899

Report message to a moderator

[Message index]

CHISQR_CVF question.

By: R.G. Stockwell on Wed, 19 August 2009 09:12

Previous Topic:	Exporting matrices without filling blank space with zeros
Next Topic:	Re: Exporting matrices without filling blank space with zeros

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Sat Nov 29 04:04:32 PST 2025

Total time taken to generate the page: 2.39560 seconds