Adaptive estimation of the stationary density of a stochastic differential equation driven by a fractional Brownian motion (Q2194048)

This paper deals with estimating the stationary density of the SDE \[ X_t = X_0 + \int_0^t b\left(X_s\right) d s + \sigma B_t^H \] where \(B^H\) is a fractional Brownian motion with Hurst parameter \(0 < H < 1\). The authors design and analyze a purely data-driven procedure based on kernel density estimation and bandwidth selection via the Goldenshluger-Lepski method. The calibration of this procedure requires new concentration inequalities proven by the authors.