Abstract:In various industrial contexts, estimating the distribution of unobserved random vectors Xi from some noisy indirect observations H(Xi) + Ui is required. If the relation between Xi and the quantity H(Xi), measured with the error Ui, is implemented by a CPU-consuming computer model H, a major practical difficulty is to perform the statistical inference with a relatively small number of runs of H. Following Fu et al. (2014), a Bayesian statistical framework is considered to make use of possible prior knowledge on the parameters of the distribution of the Xi, which is assumed Gaussian. Moreover, a Markov Chain Monte Carlo (MCMC) algorithm is carried out to estimate their posterior distribution by replacing H by a kriging metamodel build from a limited number of simulated experiments. Two heuristics, involving two different criteria to be optimized, are proposed to sequentially design these computer experiments in the limits of a given computational budget. The first criterion is a Weighted Integrated Mean Square Error (WIMSE). The second one, called Expected Conditional Divergence (ECD), developed in the spirit of the Stepwise Uncertainty Reduction (SUR) criterion, is based on the discrepancy between two consecutive approximations of the target posterior distribution. Several numerical comparisons conducted over a toy example then a motivating real case-study show that such adaptive designs can significantly outperform the classical choice of a maximin Latin Hypercube Design (LHD) of experiments. Dealing with a major concern in hydraulic engineering, a particular emphasis is placed upon the prior elicitation of the case-study, highlighting the overall feasibility of the methodology. Faster convergences and manageability considerations lead to recommend the use of the ECD criterion in practical applications.

Density-based Monte Carlo filter and its applications in estimation of unobservable variables and pharmacokinetic parameters

Transport Parameter Estimations of Plasma Transport Dynamics Using the Extended Kalman Filter

Current Statistical Model Probability Hypothesis Density Filter For Multiple Maneuvering Targets Tracking

Reinforcement Learning Based Model Selection and Parameter Estimation for Pharmacokinetic Analysis in Drug Selection.

Normalizing Flow-Based Distribution Estimation of Pharmacokinetic Parameters in Dynamic Contrast-Enhanced Magnetic Resonance Imaging.

Extended Kalman Filter for Estimation of Parameters in Nonlinear State-Space Models of Biochemical Networks.

Nonparametric estimation of a mixing distribution for a family of linear stochastic dynamical systems

Density Estimation by Monte Carlo and Quasi-Monte Carlo

PDE-Based Dynamic Density Estimation for Large-Scale Agent Systems

Online chaotic time-series prediction with the derivative-free extended Kalman filter

Comparison of Expectation-Maximization based parameter estimation using Particle Filter, Unscented and Extended Kalman Filtering techniques

Adaptive sequential Monte Carlo by means of mixture of experts

Unscented Auxiliary Particle Filter Implementation of the Cardinalized Probability Hypothesis Density Filter

Estimation of nonlinear mixed-effects continuous-time models using the continuous-discrete extended Kalman filter

Estimating States for Nonlinear Systems Using the Particle Filter

Bayesian estimation of pharmacokinetic parameters: an important component to include in the teaching of clinical pharmacokinetics and therapeutic drug monitoring

The Alive Particle Filter and Its Use in Particle Markov Chain Monte Carlo

A hybrid nonparametric multivariate density estimator with applications to risk management

Ensemble-localized Kernel Density Estimation with Applications to the Ensemble Gaussian Mixture Filter

An Outlier-Robust Kalman Filter With Adaptive Selection of Elliptically Contoured Distributions

An adaptive kriging method for solving nonlinear inverse statistical problems