I have a question regarding goodness of fit and associated probability
testing.  Here is my situation:

I have a data series (Array X) which I am performing a spectral analysis
on.  From that analysis I have found a certain number of periodic
components (which in my case will be sine waves).  I would like to test
how well the original data series (Array X) fits a known sine wave
(Array Y) of given frequency, amplitude, and phase.  I would also like
to test the statistical strength (associated probability) that the
measure of the goodness of fit could not have been achieved due to
randomness (i.e.: the probability that the measure of the goodness of
fit of Array X to Array Y could have been achieved if the data in Array
X was purely random).

Due to my lack of formal education and knowledge in this area, I am
hoping that someone here can illustrate the correct methodology for
accomplishing this.  Some tests which I have found, which may, or may
not be appropriate are:

   Pearson's correlation coefficient
   Chi Squared Goodness-of-Fit Test
   Kolmogorov-Smirnov Goodness-of-Fit Test

If you are able, would you please comment on the above tests (which I
should use and how should I use them and which should I not) and if
there are other more appropriate tests, would you please inform me.

Thank you very much in advance for your assistance.

With warmest regards.

Andrew Peskin
[EMAIL PROTECTED]


.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to