Prev: www.fashion-long-4biz.com supply nike sports shoes,nike jordan shoes,nike
Next: Multiple linear regression with volumic constraint
From: ben on 25 Jun 2010 08:49
pmtm() nicely does a multitaper PSD. and even gives confidence
intervals. thomson (1982) however describes an F-test to decide
whether there is significant energy at a particular frequency. which
of these is the preferred method for deciding on whether a signal is
present in noise?
From: Wayne King on 25 Jun 2010 09:11
ben <bja1970(a)gmail.com> wrote in message <16a04111-07db-4c81-9d53-a73bfaf9a92c(a)w31g2000yqb.googlegroups.com>...
> pmtm() nicely does a multitaper PSD. and even gives confidence
> intervals. thomson (1982) however describes an F-test to decide
> whether there is significant energy at a particular frequency. which
> of these is the preferred method for deciding on whether a signal is
> present in noise?

Hi Ben, I think if you attempting to perform a hypothesis test of whether or not the amplitude of a sine wave at a specified frequency in colored noise is zero or not, the multitaper F-test is the preferred way.

Wayne
From: ben on 28 Jun 2010 11:29
i agree. so far as i can tell matlab does not include the multitaper
F-test built-in. i know that the chronux library will do it. but
it's PSD does not adaptively weight the taper coefficients like pmtm()
does. sigh. must i really cobble this together myself? :(


> Hi Ben, I think if you attempting to perform a hypothesis test of whether or not the amplitude of a sine wave at a specified frequency in colored noise is zero or not, themultitaperF-test is the preferred way.
 | 
Pages: 1
Prev: www.fashion-long-4biz.com supply nike sports shoes,nike jordan shoes,nike
Next: Multiple linear regression with volumic constraint