Hello,

Performing correlation using FFT results in a correlation curve that is not normalized to be in the range of +/-1. Could anyone point me in the right direction on how to normalize the result?

Thanks in advance,

Easwar.

For an autocorrelation I subtract the mean of the series. For a correlation that should work as well. But I have never tried for correlation, jsut autocorrelation.

I had the same problem trying to duplication autorrelations with SAS. was able to do so only after removing mean.