Autocorrelation computes the autocorrelation function
of a signal. The autocorrelation function gives information about the
self-similarity of a signal occurring at certain lags.
Internally, a frequency-domain algorithm based on a Fast Fourier
Transform is used for calculating the discrete autocorrelation.
Warning: Note that y(0) is suppressed in order to avoid its large values dominating the result.
The autocorrelation function of a white noise signal will yield
a large peak at
=0 and nearly 0 elsewhere indicating (nearly) no
self-similarity for .
y = AutoCorr(x);
Cross correlation, Partial autocorrelation.
Oppenheim/Schafer , Conover
, Graybill , Hollander/Wolfe