Abstract
It is well-known that the linear prediction (LP) spectral estimator is equivalent to the maximum entropy spectral estimator and they are also equivalent to the maximum spectral flatness spectral estimator for AR processes of known order. The author proposes a new higher-order statistics (HOS) based linear prediction error (LPE) filter. The author also presents LP polyspectral estimator, maximum polyspectral flatness polyspectral estimator, maximum higher-order entropy polyspectral estimator and equivalencies on these polyspectral estimators. The results presented provide some theoretical foundations on the polyspectral estimation and modeling of non-Gaussian AR processes