This paper deals with the problem of the discrimination between stable and unstable time series. One criterion for the seperation is given by the size of the Lyapunov exponent, which was originally defined for deterministic systems. However, this paper will show, that the Lyapunov exponent can also be analyzed and used for ergodic stochastic time series. Experimantal results illustrate the classification by the Lyapunov exponent. Although the Lyapunov exponent is a discriminatory parameter of the asymptotic behavior and can be interpreted as a parameter of the asymptotic distribution in the stochastic case, it has to be estimated from a given time series, where the process might still be in the transient state. Experimental results will show that in special cases the estimation leads to misclassifications of the time series and the underlying process due to the uncertainty of estimators for the Lyapunov exponent.
