Centre for Mathematical Physics and Stochastics

Department of Mathematical Sciences, University of Aarhus

Funded by The Danish National Research Foundation

MPS-RR 2001-11

March 2001

The problem of estimating the parameters of a discretely observed diffusion is discussed through two one-dimensional examples. Based on simulations, the parameters are estimated using small $Delta$-optimal and other unbiased estimating functions. Small $Delta$-optimality implies that the estimation is nearly efficient when the discrete observations are close together in time, and this effect is clearly visible from the simulations. It is also seen that the small $Delta$-optimal estimating functions perform well, even when the observations are not close together, and that they are quite robust when the true observations are replaced by rounded ones.

Availability: [ gzipped `ps`

-file ] [ `pdf`

-file ]