Parameter estimation in nonlinear stochastic differential equations

We discuss the problem of parameter estimation in nonlinear stochastic differential equations based on sampled time series. A central message from the theory of integrating stochastic differential equations is that there exists in general two time scales, i.e. that of integrating these equations and that of sampling. We argue that therefore maximum likelihood estimation is computational extremely expensive. We discuss the relation between maximum likelihood and quasi maximum likelihood estimation. In a simulation study, we compare the quasi maximum likelihood method with an approach for parameter estimation in nonlinear stochastic differential equations that disregards the existence of the two time scales.