Posted by Don Vaught on July 16, 1998 at 22:23:36:

In Reply to: Re: Bivariate Normal Distribution posted by K. M. Anirudh on February 14, 1998 at 11:51:45:

I noticed you wanted a numerical algorithm for

computing the cdf of a bivariate normal. I found a

method in an article. The reference is:

Z. Drezner, "Computation of the Bivariate Normal

Integral," Mathematics of Computation, 32 (January

1978), P. 277-279.

I first found it in:

John Hull, Options,Futures,and other Derivative

Securities, Prentice Hall, 93.

I hope it works out.

: Richard Allen,

: : I am trying to find an algorithm to calulate

: : accurately the cumulative density function for the

: : bivariate normal distribution. Any ideas?

: What is the difficulty in implementing the old trapezoidal

: rule with very small increment? If you are looking for the

: most efficient one it may not be useful. I used the same

: method recently to integrate the desnsity of a Hotelling's

: density for sample correlation coefficient, primarily

: because I had no reference book in hand that gave me

: the distribution function readymade and it was a week-end.

: My problem was simpler in that it was univariate, but I think

: it can be used for the bivariate case, if you don't

: mind the CPU usage.

: Hope you would find a better solution:-)

:

: -- Anirudh