18
0

Measure estimation on a manifold explored by a diffusion process

Abstract

From the observation of a diffusion path (Xt)t[0,T](X_t)_{t\in [0,T]} on a compact connected dd-dimensional manifold MM without boundary, we consider the problem of estimating the stationary measure μ\mu of the process. Wang and Zhu (2023) showed that for the Wasserstein metric W2W_2 and for d5d\ge 5, the convergence rate of T1/(d2)T^{-1/(d-2)} is attained by the occupation measure of the path (Xt)t[0,T](X_t)_{t\in [0,T]} when (Xt)t[0,T](X_t)_{t\in [0,T]} is a Langevin diffusion. We extend their result in several directions. First, we show that the rate of convergence holds for a large class of diffusion paths, whose generators are uniformly elliptic. Second, the regularity of the density pp of the stationary measure μ\mu with respect to the volume measure of MM can be leveraged to obtain faster estimators: when pp belongs to a Sobolev space of order >0\ell>0, smoothing the occupation measure by convolution with a kernel yields an estimator whose rate of convergence is of order T(+1)/(2+d2)T^{-(\ell+1)/(2\ell+d-2)}. We further show that this rate is the minimax rate of estimation for this problem.

View on arXiv
Comments on this paper