Nonparametric statistical inference for drift vector fields of multi-dimensional diffusions

The problem of determining a periodic Lipschitz vector field from an observed trajectory of the solution of the multi-dimensional stochastic differential equation \begin{equation*} dX_t = b(X_t)dt + dW_t, \quad t \geq 0, \end{equation*} where is a standard -dimensional Brownian motion, is considered. Convergence rates of a penalised least squares estimator, which equals the maximum a posteriori (MAP) estimate corresponding to a high-dimensional Gaussian product prior, are derived. These results are deduced from corresponding contraction rates for the associated posterior distributions. The rates obtained are optimal up to log-factors in -loss in any dimension, and also for supremum norm loss when . Further, when , nonparametric Bernstein-von Mises theorems are proved for the posterior distributions of . From this we deduce functional central limit theorems for the implied estimators of the invariant measure . The limiting Gaussian process distributions have a covariance structure that is asymptotically optimal from an information-theoretic point of view.
View on arXiv