Statistics for Discrete-time nonlinear filtering algorithms using Gauss-Hermite quadrature: new computationally efficient methods are proposed for more accurately analyzing and modeling dynamic processes that are nonlinear and subject to non-Gaussian noise