The Dynamic Correlation and Predictability of Stochastic Dynamic Systems in the Presence of Fluctuational Forces

The problem of predicting the processes induced by strange attractors in the presence of 'measurement' noise and internal fluctuation forces is studied. It is shown that stochastic systems admit of prediction over a finite time, called the dynamic memory time. It is important to note that this time may appreciably exceed the autocorrelation time. To elucidate the dynamic nature of the random processes, it is suggested that the concept of the degree of dynamic correlation be used; this is defined as the coefficient of mutual correlation between the studied and predicted processes.