Abstract
An algorithm for the linear least squares prediction of some classes of non‐stationary processes is obtained. The non‐stationary models under study recover and unify, in a single framework, the Kolmogorov‐Wiener and the Kalman theory of prediction, and also contain other non-stationary classes such as second order martingale difference processes.
Present Address: Department of Mathematics and System Research Center, University of Maryland
Research Supported by the ONR Contract No. N00014 86C 0227 and in Part by the AFSOR Contract No. F49620 85C 0144
Present Address: Department of Mathematics and System Research Center, University of Maryland
Research Supported by the ONR Contract No. N00014 86C 0227 and in Part by the AFSOR Contract No. F49620 85C 0144
Notes
Present Address: Department of Mathematics and System Research Center, University of Maryland
Research Supported by the ONR Contract No. N00014 86C 0227 and in Part by the AFSOR Contract No. F49620 85C 0144