GaussianQuadrature and truncated normal
- To: mathgroup at smc.vnet.net
- Subject: [mg19492] GaussianQuadrature and truncated normal
- From: "H. J. Wang" <hjwang at ieas.econ.sinica.edu.tw>
- Date: Sat, 28 Aug 1999 15:53:15 -0400
- Organization: Computing Center, Academia Sinica
- Sender: owner-wri-mathgroup at wolfram.com
Hi,
Does anyone know how to do a numerical integration of a
truncated normal distribution using Gaussian quadratures?
I have a function f(x,z), where x is an unknown parameter,
and z has a truncated normal probability distribution with E(z) = 0.
The maximum and minimum values of z are zmax and zmin, respectively. I
want to integrate f(x,z) over zmax and zmin, and express the result as
a finite sum of the `f()' function: SUM_i f(x, g_i)*w_i. But I guess
I cannot take the nods and weights (ie, {g_i, w_i}) directly from
Gaussian quadratures because z is NOT a simple normal but a truncated
normal. Am I correct? If so, how should I proceed?
Any advice will be greatly appreciated.