Nonparametric estimation of homogenized invariant measures from multiscale data via Hermite expansion
Preprint
Abstract
We consider the problem of density estimation in the context of multiscale Langevin diffusion processes, where a single-scale homogenized surrogate model can be derived. In particular, our aim is to learn the density of the invariant measure of the homogenized dynamics from a continuous-time trajectory generated by the full multiscale system. We propose a spectral method based on a truncated Fourier expansion with Hermite functions as orthonormal basis. The Fourier coefficients are computed directly from the data owing to the ergodic theorem. We prove that the resulting density estimator is robust and converges to the invariant density of the homogenized model as the scale separation parameter vanishes, provided the time horizon and the number of Fourier modes are suitably chosen in relation to the multiscale parameter. The accuracy and reliability of this methodology is further demonstrated through a series of numerical experiments.Keywords
All Related Versions
This publication has 23 references indexed in Scilit:
- Limit Theorems for One-Dimensional Homogenized Diffusion ProcessesJournal of Statistical Physics, 2026
- Nonparametric moment method for scalar McKean–Vlasov stochastic differential equationsESAIM: Probability and Statistics, 2025
- Filtered data based estimators for stochastic processes driven by colored noiseStochastic Processes and their Applications, 2024
- Stochastic gradient descent in continuous time for drift identification in multiscale diffusionsESAIM: Mathematical Modelling and Numerical Analysis, 2024
- Eigenfunction martingale estimating functions and filtered data for drift estimation of discretely observed multiscale diffusionsStatistics and Computing, 2022
- Optimal convergence rates for the invariant density estimation of jump-diffusion processesESAIM: Probability and Statistics, 2022
- Drift Estimation of Multiscale Diffusions Based on Filtered DataFoundations of Computational Mathematics, 2021
- A new framework for extracting coarse-grained models from time series with multiscale structureJournal of Computational Physics, 2015
- Parameter Estimation for Multiscale DiffusionsJournal of Statistical Physics, 2007
- New orthogonality relations for the Hermite polynomials and related Hilbert spacesJournal of Mathematical Analysis and Applications, 1990