Discrimination theory of physical processes

Abstract

Discrimination problems of stochastic processes and time-evolving physical states are treated on the basis of an information theoretical method. From a set of received data we can select a suitably assumed probability distribution. The basic method is applied to discriminate colored stochastic processes of the Brownian motion and laser system with nonlinearity and fluctuations. Dynamical state change is clearly specified: This is quite difficult by a direct observation of the data. The methodology is further extended to include quantum mechanical systems.