On the difference equation for statistical estimation and forecasting

A straightforward computing algorithm is developed for statistical estimation and forecasting the parameters of the polynomial deterministic basis of random processes by the method of least squares. A procedure is described whereby a linear difference equation with variable coefficients is obtained which is the core of the algorithm.