Simulation of truncated normal variables

By Christian P. Robert. 2009

570 citations recorded by the source corpus.

We provide simulation algorithms for one-sided and two-sided truncated normal distributions. These algorithms are then used to simulate multivariate normal variables with convex restricted parameter space for any covariance structure.

Subfields: Artificial Intelligence, Management Science and Operations Research, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/88b51eb03f1b0a3449136bbb16b8aa40ed0c1307

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 Management Science and Operations Research

In Statistics and Probability

Citations in the corpus

Consistent inference in fixed-effects stochastic frontier models (2017, methodology): “The simulation of inefficiency vector ui is performed using the MCMC approach outlined in Geweke (1991) or in Robert (1995), which uses a Gibbs algorithm for sampling from an arbitrary multivariate truncated normal distribution.”
Censored time series analysis with autoregressive moving average models (2007, unknown): “the matrix § with j-th row and column deleted (see Robert, 1995 ).”
Efficient sampling methods for truncated multivariate normal and student-t distributions subject to linear inequality constraints (2015, methodology): “Geweke (1991) and Robert (1995) develop the sampling methods based on accept-reject algorithm.”
Bayesian auxiliary variable models for binary and multinomial regression (2006, background): “The full conditional for each element z i is then truncated normal, which is straightforward to sample from, see for example Robert (1995).”
Gene selection: a Bayesian variable selection approach (2003, background): “cDNA microarrays provide expression measurements for thousands of genes at once (Duggan et al., 1999; Schena et al., 1995).”
MCMCglmm Course Notes (2012, background): “TECHNICAL DETAILS 119 yi = li and zero otherwise so li = yi with out the need for updating, ii) when yi is discrete and modelled using family="threshold" then Equation defines a truncated normal distribution and can be slice sampled (Robert, 1995) and iii) when yi is missing fi(yi|li) is not defined”
Generalized Nonlinear Modeling With Multivariate Free-Knot Regression Splines (2003, background): “zijZ¡i;Y; B; 6; ̄/ is a truncated normal distribution (18) for which ef cient sampling algorithms exist (Robert 1995).”
Estimation of Errors in the Inverse Modeling of Accidental Release of Atmospheric Pollutant: Application to the Reconstruction of the Cesium-137 and Iodine-131 Source Terms from the Fukushima Daiichi Power Plant (2012, methodology): “First, a1 is computed and u1 is drawn from T N ð0; 1;a1;þ∞Þ , following the exponential accept-reject algorithm proposed by Robert [1995] if a1 > 0.”
MCMC Estimation of Restricted Covariance Matrices (2009, background): “This is due, on the one hand, to the fact that binary data are less informative about covariate effects because the threshold-crossing nature of the discretization transformation yij = 1{y∗ ij > 0} only contains information on signs, and not on magnitudes; on the other hand, the larger inefficiency ”
Factorization Machines with libFM (2012, background): “(39) N ( y(xi,8),1) d(ypi = 0), if yi has the positive class Sampling from this distribution is ef.cient [Robert 1995].”
Learning about physical parameters: the importance of model discrepancy (2014, methodology): “Operationally, if the mean μ̃q,j is negative we sample the normal distribution until we get a negative number, otherwise we use the rejection sampling algorithm of Robert (1995).”