Likelihood based statistics for partially observed diffusion processes
Abstract
The purpose of this paper is to study some statistical problems: parameter estimation, binary detection, change detection (disorder problem), etc. for partially observed diffusion processes, using the likelihood approach. It is shown that the stochastic PDE related to the state estimation problem, provides also a way to compute the likelihood function/ratio. A recent result on consistency of the MLE, in the small noise asymptotics, is also presented.