Abstract:
A nonlinear algorithm is proposed for solution of the problem of extrapolating a random process in a computer by statistical modeling which minimizes the mean square prediction error. Some results from an experimental study of the properties of the algorithm are given.