Bayesian computing with INLA: New features

By Thiago G. Martins, Daniel P. Simpson, Finn Lindgren, Håvard Rue. Computational Statistics & Data Analysis, 2012

626 citations recorded by the source corpus.

The INLA approach for approximate Bayesian inference for latent Gaussian models has been shown to give fast and accurate estimates of posterior marginals and also to be a valuable tool in practice via the R-package R-INLA. New developments in the R-INLA are formalized and it is shown how these features greatly extend the scope of models that can be analyzed by this interface. The current default method in R-INLA to approximate the posterior marginals of the hyperparameters using only a modest number of evaluations of the joint posterior distribution of the hyperparameters, without any need for numerical integration, is discussed.

Subfields: Artificial Intelligence, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/f3257d04ac5b73fa15eeeab93827b1abc3ed23e7

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Statistics and Probability

Citations in the corpus

Robustly detecting differential expression in RNA sequencing data using observation weights (2013, methodology): “ShrinkBayes introduces the ZINB and performs inference using integrated nested Laplace approximations (INLA) (30,31) and provides Bayesian FDR and local false discovery rate (lfdr) (32) estimates.”
Bayesian Computing with INLA: A Review (2016, methodology): “Here, we give a short summary of these, for more details see Martins et al. (2013). replicate Each model component given as a f() -term can be replicated, creating nrep iid replications with shared hyperparameters.”
Fast Calibrated Additive Quantile Regression (2017, methodology): “The first proposal is implemented in the INLA software (Martins et al. 2013), but the associated documentation discourages its use.”
A Unified Framework for Fitting Bayesian Semiparametric Models to Arbitrarily Censored Survival Data, Including Spatially Referenced Data (2017, positive): “BayesX and R-INLA (Martins et al. 2013) both include the Matérn correlation function ρ(s, s′;φ) = [2ν−1 (ν)]−1(φ‖s − s′‖)νKν (φ‖s − s′‖)whereKν is the modified Bessel function of order ν. Noting weakly identified Matérn parameters, BayesX requires the user to fix one of ν ∈ {0.5, 1.5, 2.5, 3.5} (ν =”
Programming With Models: Writing Statistical Algorithms for General Model Structures With NIMBLE (2015, methodology): “For example, lme4 (Bates et al., 2014), MCMCglmm (Hadﬁeld, 2010), R-INLA (Martins et al., 2013), and others each use a diﬀerent system for GLMM speciﬁcation.”
Data Integration for Large-Scale Models of Species Distributions. (2020, methodology): “These approximations have been combined with efficient methods for approximating the continuous space in the INLA software [67,68], which makes it possible to fit many complex ecological models efficiently using point”
Bayesian Spatial Modelling with R-INLA (2015, methodology): “For further details on the computational and inferential methods in the R-INLA package we refer to Martins, Simpson, Lindgren, and Rue (2013).”
Bayesian analysis of measurement error models using integrated nested Laplace approximations (2015, result): “This is in contrast with other applications where the respective coefficient is often fixed to 1 (Martins et al., 2013).”
Penalising Model Component Complexity: A Principled, Practical Approach to Constructing Priors (2014, methodology): “As developers of the R-INLA (see http: //www.r-inla.org/) package, which performs approximate Bayesian inference for latent Gaussian models (Rue et al., 2009; Lindgren et al., 2011; Martins et al., 2013), we are left with two unpalatable choices.”