Modeling of a nonstationary stochastic process produced by a linear time-varying system

We have clarified in this work the conditions for a discrete-time Gaussian process to be an auto-regressive (AR-), moving average (MA-), or ARMA-process based on the canonical representation of the Gaussian process. We have already established the algorithm for synthesizing circuits representing the processes, and also obtained basic theorems on the convergence properties of the algorithms of linear prediction of the processes. These results will be published in a following paper.