Parameter Estimation with Data-Driven Nonparametric Likelihood Functions

@article{Jiang2019ParameterEW,
  title={Parameter Estimation with Data-Driven Nonparametric Likelihood Functions},
  author={Shixiao W. Jiang and John Harlim},
  journal={Entropy},
  year={2019},
  volume={21}
}
In this paper, we consider a surrogate modeling approach using a data-driven nonparametric likelihood function constructed on a manifold on which the data lie (or to which they are close). The proposed method represents the likelihood function using a spectral expansion formulation known as the kernel embedding of the conditional distribution. To respect the geometry of the data, we employ this spectral expansion using a set of data-driven basis functions obtained from the diffusion maps… 

Figures from this paper

Modeling of missing dynamical systems: deriving parametric models using a nonparametric framework

TLDR
This paper considers modeling missing dynamics with a nonparametric non-Markovian model, constructed using the theory of kernel embedding of conditional distributions on appropriate reproducing kernel Hilbert spaces (RKHS), equipped with orthonormal basis functions, and shows that the success of various parametric modeling approaches that were proposed in various domains of applications can be understood through the RKHS representations.

Kernel Embedding Linear Response

TLDR
A theoretical guarantee for the convergence of the estimator to the underlying actual linear response statistics and a statistical error bound for the density estimation that accounts for the Monte-Carlo averaging over non-i.i.d time series and biases due to finite basis truncation are provided.

Estimating linear response statistics using orthogonal polynomials: An rkhs formulation

We study the problem of estimating linear response statistics under external perturbations using time series of unperturbed dynamics. Based on the fluctuation-dissipation theory, this problem is

A parameter estimation method using linear response statistics: Numerical scheme.

TLDR
From the numerical experiments, it is found that the proposed approach is superior compared to the conventional approach that uses equilibrium statistics to determine the parameters.

Linear Response Based Parameter Estimation in the Presence of Model Error

Joint state-parameter estimation of a nonlinear stochastic energy balance model from sparse noisy data

TLDR
This work investigates joint parameter-state estimation by the regularized posterior in a physically motivated nonlinear stochastic energy balance model (SEBM) for paleoclimate reconstruction by normalizing the likelihood and by imposing physical constraints through priors of the parameters and states.

Bayesian Instability of Optical Imaging: Ill Conditioning of Inverse Linear and Nonlinear Radiative Transfer Equation in the Fluid Regime

For the inverse problem in physical models, one measures the solution and infers the model parameters using information from the collected data. Oftentimes, these data are inadequate and render the

References

SHOWING 1-10 OF 44 REFERENCES

A parameter estimation method using linear response statistics: Numerical scheme.

TLDR
From the numerical experiments, it is found that the proposed approach is superior compared to the conventional approach that uses equilibrium statistics to determine the parameters.

Stochastic spectral methods for efficient Bayesian solution of inverse problems

The Bayesian Approach to Inverse Problems

These lecture notes highlight the mathematical and computational structure relating to the formulation of, and development of algorithms for, the Bayesian approach to inverse problems in

Data-Driven Computational Methods

TLDR
This book describes two classes of computational methods to leverage data for modeling dynamical systems, concerned with data fitting algorithms to estimate parameters in parametric models that are postulated on the basis of physical or dynamical laws.

A stochastic collocation approach to Bayesian inference in inverse problems

We present an efficient numerical strategy for the Bayesian solution of inverse problems. Stochastic collocation methods, based on generalized polynomial chaos (gPC), are used to construct a

Spectral likelihood expansions for Bayesian inference

Diffusion Forecasting Model with Basis Functions from QR-Decomposition

TLDR
The diffusion forecasting is a nonparametric approach that provably solves the Fokker–Planck PDE corresponding to Itô diffusion without knowing the underlying equation and a new set of basis functions constructed by orthonormalizing selected columns of the diffusion matrix and its leading eigenvectors is proposed.

Variable Bandwidth Diffusion Kernels

Correcting Biased Observation Model Error in Data Assimilation

AbstractWhile the formulation of most data assimilation schemes assumes an unbiased observation model error, in real applications model error with nontrivial biases is unavoidable. A practical