• Corpus ID: 245650273

Frequentist perspective on robust parameter estimation using the ensemble Kalman filter

@article{Reich2022FrequentistPO,
  title={Frequentist perspective on robust parameter estimation using the ensemble Kalman filter},
  author={Sebastian Reich},
  journal={ArXiv},
  year={2022},
  volume={abs/2201.00611}
}
  • S. Reich
  • Published 3 January 2022
  • Mathematics
  • ArXiv
. Standard maximum likelihood or Bayesian approaches to parameter estimation for stochastic differential equations are not robust to perturbations in the continuous-in-time data. In this paper, we give a rather elementary explanation of this observation in the context of continuous-time parameter estimation using an ensemble Kalman filter. We employ the frequentist perspective to shed new light on two robust estimation techniques; namely subsampling the data and rough path corrections. We… 

Figures from this paper

Dimension free non-asymptotic bounds on the accuracy of high dimensional Laplace approximation

TLDR
The classical results on Laplace approximation in a modern non-asymptotic and dimension free form are revisited and the issue of using a Gaussian approximation with inexact parameters with the focus on replacing the Maximum a Posteriori (MAP) value by the posterior mean is addressed.

Rough McKean-Vlasov dynamics for robust ensemble Kalman filtering

TLDR
A McKean-Vlasov equation that contains the data stream as a common driving rough path is studied, establishing propagation of chaos for the associated interacting particle system, which is suggestive of a numerical scheme that can be viewed as an extension of the ensemble Kalman filter to a rough-path framework.

References

SHOWING 1-10 OF 35 REFERENCES

and s

Physical Brownian motion in a magnetic field as a rough path

The indefinite integral of the homogenized Ornstein-Uhlenbeck process is a well-known model for physical Brownian motion, modelling the behaviour of an object subject to random impulses [L. S.

Differential Equations Driven by Rough Paths: An Approach via Discrete Approximation

driving path x(t) is nondifferentiable, has recently been developed by Lyons. I develop an alternative approach to this theory, using (modified) Euler approximations, and investigate its

Maximum likelihood estimation for multiscale diffusions

  • Stochastic Processes and their Applications, 19:3173–3210
  • 2009

State and Parameter Estimation from Observed Signal Increments

TLDR
This paper takes an appropriate class of McKean–Vlasov equations as the starting point to derive ensemble Kalman–Bucy filter algorithms for combined state and parameter estimation for continuous-time data assimilation where the model and measurement errors are correlated and both states and parameters need to be identified.

Rough McKean-Vlasov dynamics for robust ensemble Kalman filtering

TLDR
A McKean-Vlasov equation that contains the data stream as a common driving rough path is studied, establishing propagation of chaos for the associated interacting particle system, which is suggestive of a numerical scheme that can be viewed as an extension of the ensemble Kalman filter to a rough-path framework.

L2 convergence of smooth approximations of Stochastic Differential Equations with unbounded coefficients

TLDR
This paper considers convergence in mean in the uniform topology of piecewise linear approximations of Stochastic Differential Equations with C^1 drift and C^2 diffusion coefficients with uniformly bounded derivatives using a combination of traditional stochastic analysis and rough path techniques.

Drift Estimation of Multiscale Diffusions Based on Filtered Data

TLDR
This work avoids subsampling and works instead with filtered data, found by application of an appropriate kernel function, and compute maximum likelihood estimators based on the filtered process, which are asymptotically unbiased and demonstrate numerically the advantages of the method with respect to subsampled.

Posterior contraction rates for non-parametric state and drift estimation

We consider a combined state and drift estimation problem for the linear stochastic heat equation. The infinite-dimensional Bayesian inference problem is formulated in terms of the Kalman-Bucy filter