Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
st: gausshermite question
From
"David Rivenbark" <[email protected]>
To
<[email protected]>
Subject
st: gausshermite question
Date
Sat, 29 Mar 2014 14:44:53 -0400
I'm using -gausshermite- to approximate the expected value of a
variable 0 <= z <= 1 that is distributed logit-normal with mean 0.4528 and
variance 0.16. I have altered the original ADO file to allow me to set more
than 100 observations, thus more than 98 nodes.
With 150 nodes -myghermite- retuned approximately 0.60 as it should
(Frederic and Lad 2008).
. myghermite 1/(1+exp(-x)) , sigma(.4) mu(.4528) d n(150) observations(152)
int_R (1/(1+exp(-x)))g(x/sigma=.4)dx= 0.60735817
However, when I set 200 nodes -myhermite- returns 0.
. myghermite 1/(1+exp(-x)) , sigma(.4) mu(.4528) d n(200) observations(202)
int_R (1/(1+exp(-x)))g(x/sigma=.4)dx= 0.00000000
What is causing the zero value? Is there a way to fix this issue?
Reference:
Patrizio Frederic & Frank Lad (2008) Two Moments of the Logitnormal
Distribution, Communications in Statistics - Simulation and Computation,
37:7, 1263-1269
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/faqs/resources/statalist-faq/
* http://www.ats.ucla.edu/stat/stata/