An ARMA bispectrum estimation by two AR estimation process approach

Shows a new approach of the bispectrum estimation. The authors obtained not only a good estimation for bispectrum, but also the best order of the estimation model. This method is applied to the system identification of the nonGaussian process.<
>