Thank to Kit Baum, the -gausshermite- module has been added to SSC.
-gausshermite- allow to estimate the integrals of the form
f(x)g(x/mu,sigma) over all the reals by Gauss-Hermite quadrature, where
g(x/mu,sigma) is the gaussian distribution function of mean mu and
variance sigma^2. It allow to estimate E(f(X)) if X is distributed among
g(X/mu,sigma).
You must describe the f(x) function, and -optionnally- the values of mu
and sigma. f(x) is a function who depend only of x. The module return
only one value (this module is closer of a Stata function than an usual
command because all the used information is contained in the written of
the command, and so the result is the same, whatever the current dataset).
Example :
. gausshermite x^2 , sigma(10) mu(5)
if f(x)=x^2, mu=5 and sigma=10
The estimated value of the integral is save in r(int). You must display
the value in using the -display- option,
for example
. gausshermite x^2 , sigma(10) mu(5) display
will return :
int_R (x^2)g(x/sigma=10)dx=125.00001203
The number of quadrature nodes can be define with the nodes option.
By default, sigma=1, mu=0 and nodes=12
The module use the -ghquadm- module to estimate the quadrature nodes
and weights. This program can be download in writting -findit ghquadm-
(in STB-61).
Jean-Benoit Hardouin
--
***************************************************************************
Jean-Benoit Hardouin
Biostatisticien
Observatoire R�gional de la Sant� du Centre
BP 2439
1, rue Porte Madeleine
45032 Orl�ans Cedex 1
t�l : 02 38 74 48 80
fax : 02 38 74 48 81
Email : [email protected]
**************************************************************************
