Nonlinear orthogonal parametrization and modeling of higher-order non-Gaussian time-series

The nonlinear (second-degree) generalization of the linear J-orthogonal Schur parametrization/innovations and orthogonal stochastic modeling of second-order time-series to higher-(fourth-) order non-Gaussian time-series is proposed. The J-orthogonal realization of the nonlinear parametrization/innovations filter, and the orthogonal realization of the nonlinear modeling filter is presented.