AN APPROACH TO STOCHASTIC SYSTEMS MODELLING

Abstract

The problem of modelling stochastic systems when only a partial statistical description is available is considered. Specifically, a procedure is proposed for assigning an optimal joint probability model relating the input and output of the system where the partial statistical description becomes constraints. The Mutual Information functional is used to establish the model leading to a criteria which is optimal in an information theory sense. Results showing general solutions for cases of interest in digital communications as well as continuous systems with noise variance knowledge are given.

Citation

NOONAN, J.P. and MARCUS, J.R. (1986), "AN APPROACH TO STOCHASTIC SYSTEMS MODELLING", Kybernetes, Vol. 15 No. 4, pp. 225-229. https://doi.org/10.1108/eb005744