MINIMUM CONTRAST ESTIMATION FOR DISCRETELY OBSERVED DIFFUSION PROCESSES WITH SMALL DISPERSION PARAMETER

摘要

The parametric estimation of both drift and diffusion coefficient parameters for d-dimensional diffusion processes with small dispersion parameter e is stated when the data are discretely observed at equidistant time points k, k = 0, 1,...,n. Using the contrast function based on a Gaussian approximation to the transition density, we present asymptotic properties for the minimum contrast estimator as e tends to 0 and n tends to∞simultaneously.