A Mata Geweke–Hajivassiliou–Keane multivariate normal simulator

An accurate and efficient numerical approximation of the multivariate normal (MVN) distribution function is necessary for obtaining maximum likelihood estimates for models involving the MVN distribution. Numerical integration through simulation (Monte Carlo) or number-theoretic (quasi–Monte Carlo) techniques is one way to accomplish this task. One popular simulation technique is the Geweke–Hajivassiliou–Keane MVN simulator. This paper reviews this technique and introduces a Mata function that implements it. It also computes analytical first-order derivatives of the simulated probability with respect to the variables and the variance–covariance parameters.


Issue Date:
2006
Publication Type:
Journal Article
DOI and Other Identifiers:
st0102 (Other)
PURL Identifier:
http://purl.umn.edu/117569
Published in:
Stata Journal, Volume 06, Number 2
Page range:
190-213
Total Pages:
24

Record appears in:



 Record created 2017-04-01, last modified 2017-11-24

Fulltext:
Download fulltext
PDF

Rate this document:

Rate this document:
1
2
3
 
(Not yet reviewed)