Inference on high-dimensional implicit dynamic models using a guided intermediate resampling filter

DOI10.1007/S11222-020-09957-3zbMATH Open1452.62598arXiv1708.08543OpenAlexW3037372288MaRDI QIDQ81239FDOQ81239

Authors: Joonha Park, Joonha Park, Edward L. Ionides, Edward L. Ionides 

Publication date: 26 June 2020

Published in: Statistics and Computing (Search for Journal in Brave)

Abstract: We propose a method for inference on moderately high-dimensional, nonlinear, non-Gaussian, partially observed Markov process models for which the transition density is not analytically tractable. Markov processes with intractable transition densities arise in models defined implicitly by simulation algorithms. Widely used particle filter methods are applicable to nonlinear, non-Gaussian models but suffer from the curse of dimensionality. Improved scalability is provided by ensemble Kalman filter methods, but these are inappropriate for highly nonlinear and non-Gaussian models. We propose a particle filter method having improved practical and theoretical scalability with respect to the model dimension. This method is applicable to implicitly defined models having analytically intractable transition densities. Our method is developed based on the assumption that the latent process is defined in continuous time and that a simulator of this latent process is available. In this method, particles are propagated at intermediate time intervals between observations and are resampled based on a forecast likelihood of future observations. We combine this particle filter with parameter estimation methodology to enable likelihood-based inference for highly nonlinear spatiotemporal systems. We demonstrate our methodology on a stochastic Lorenz 96 model and a model for the population dynamics of infectious diseases in a network of linked regions.

Full work available at URL: https://arxiv.org/abs/1708.08543

Recommendations

• Ensemble Kalman methods for high-dimensional hierarchical dynamic space-time models
• A stable particle filter for a class of high-dimensional state-space models
• Iterated filtering
• Implicit particle filters for data assimilation
• Nonparametric particle filtering approaches for identification and inference in nonlinear state-space dynamic systems

zbMATH Keywords

sequential Monte Carlocurse of dimensionalityimplicit modelsparticle filterplug-and-play propertyspatiotemporal inference

Mathematics Subject Classification ID

Directional data; spatial statistics (62H11) Applications of statistics to biology and medical sciences; meta analysis (62P10) Monte Carlo methods (65C05) Inference from stochastic processes and prediction (62M20) Sequential estimation (62L12)

Cites Work

• Sequential Monte Carlo Methods in Practice
• Title not available (Why is that?)
• Time series analysis via mechanistic models
• Stability and uniform approximation of nonlinear filters using the Hilbert metric and application to particle filters
• A comparison of inferential methods for highly nonlinear state space models in ecology and epidemiology
• Nonlinear data assimilation
• Efficient data assimilation for spatiotemporal chaos: a local ensemble transform Kalman filter
• Central limit theorem for sequential Monte Carlo methods and its application to Bayesian inference
• Twisted particle filters
• On the stability of sequential Monte Carlo methods in high dimensions
• Lookahead strategies for sequential Monte Carlo
• Adaptive joint detection and decoding in flat-fading channels via mixture Kalman filtering
• Title not available (Why is that?)
• A nonparametric ensemble transform method for Bayesian inference
• Inference for dynamic and latent variable models via iterated perturbed Bayes maps
• Title not available (Why is that?)
• Sequential Monte Carlo with Highly Informative Observations
• Curse-of-dimensionality revisited: Collapse of the particle filter in very large scale systems
• Stability properties of some particle filters
• A random map implementation of implicit filters
• Sequential Imputations and Bayesian Missing Data Problems
• Title not available (Why is that?)
• Filtering via Simulation: Auxiliary Particle Filters
• Random-walk models of network formation and sequential Monte Carlo methods for graphs
• Particle Markov Chain Monte Carlo Methods
• Title not available (Why is that?)
• Blind Deconvolution via Sequential Imputations
• Nonasymptotic analysis of adaptive and annealed Feynman-Kac particle models
• Mathematical statistics. Basic ideas and selected topics. Volume I
• Title not available (Why is that?)
• A stable particle filter for a class of high-dimensional state-space models
• On generating Monte Carlo samples of continuous diffusion bridges
• Efficient implementation of Markov chain Monte Carlo when using an unbiased likelihood estimator
• Second-order Accurate Ensemble Transform Particle Filters
• Sequential Monte Carlo without likelihoods
• Error bounds and normalising constants for sequential Monte Carlo samplers in high dimensions
• SMC2: An Efficient Algorithm for Sequential Analysis of State Space Models
• On the stability of interacting processes with applications to filtering and genetic algorithms
• On parallel implementation of sequential Monte Carlo methods: the island particle model
• Scalable inference for Markov processes with intractable likelihoods
• Can local particle filters beat the curse of dimensionality?

Cited In (7)

• An iterated block particle filter for inference on coupled dynamic systems with shared and unit-specific parameters
• Sequential estimation of temporally evolving latent space network models
• A comparison of inferential methods for highly nonlinear state space models in ecology and epidemiology
• Title not available (Why is that?)
• Ensemble Kalman methods for high-dimensional hierarchical dynamic space-time models
• Stratified epidemic model using a latent marked Hawkes process
• Bagged Filters for Partially Observed Interacting Systems

Uses Software

• pomp

This page was built for publication: Inference on high-dimensional implicit dynamic models using a guided intermediate resampling filter

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q81239)