cum4.m

Computes the 3 dimensional 4th order cumulant biased estimate from an input signal matrix & lag.

cum4.m computes the 3 dimensional 4th order cumulant biased estimate from an input signal matrix and a maximum lag value. As an example, a 256 sample signal was computed to +/- 31 lags in approximately 50 seconds on a 200 MHz Pentium Pro system, according to:

>> x=(rand(1,256)-0.5)*20;
>> y=cum4(x,31);

cum4.m employs the Signal Processing Toolbox 4.1 function xcorr.m.