Computing Probability Integrals of a Bivariate Normal Distribution

by Saeed Maghsoodloo and Chien-Lung Huang.

Abstract: The bivariate normal integral has been computed and tabulated in several sources in thepast 40 years. The objective of this paper is two-fold. First, matrix algebra and geometry is usedto reduce the complexity of evaluating the normal double integral thru a transformation to asingle integral of the cumulative univariate standardized normal distribution. As a result, theaccuracy for the volume under the bivariate normal density is at least through 7 decimals.Secondly, the objective is to present a very user-friendly software that computes the volume overboth elliptical and rectangular regions. The user may specify the level of accuracy by inputtingthe number of subintervals for the Simpson's Rule.

Key Words: Bivariate Normal Integral, Matrix Transformation, Computer Software

Authors:
Saeed Maghsoodloo, maghsod@eng.auburn.edu
Chien-Lung Huang

Editor: John Hinde, jph@msor0.ex.ac.uk

READING THE ARTICLE: You can read the article in portable document (.pdf) format (620132 bytes.)

NOTE: The content of this article is the intellectual property of the authors, who retains all rights to future publication.

This page has been accessed 3574 times since July 24, 2006.


Return to the InterStat Home Page.