Search All of the Math Forum:
Views expressed in these public forums are not endorsed by
NCTM or The Math Forum.



Re: aprroximation to function of a statistitics
Posted:
Aug 14, 2012 11:18 AM


On 20120814, oercim <oercim@yahoo.com> wrote:
> Hello I need help, I want to calculate variance of functional. Let v1=var(x) > My aim is to calculate a function of x depending parameter "a":
> var(f(x,a)) where "a" is the parameter.
> To approximate this variance I used delta method. However in simulation results for some values of "a" the approximation didn't work well possibly heavy nonlinearity of g.
> I dont want to use resampling mathods(like bootstrap or jacknife). Is there another ways to approximate the variance. Is higher order approximation possible? I searched google but I could not find a sufficient source. If you help me about this object, I will be very glad.
You might be able to expand the expression, but you will need more than the variance to carry out the calculations.
One way to see if nonlinearity is a problem is to check the means in the simulation. This checks for quadratic terms, but the means behaving is not a guarantee. Another possible source of the observed problem is the pseudorandom numbers being used; pseudorandom numbers are NOT random, and there are ways around that problem by incorporating physical random numbers.
 This address is for information only. I do not claim that these views are those of the Statistics Department or of Purdue University. Herman Rubin, Department of Statistics, Purdue University hrubin@stat.purdue.edu Phone: (765)4946054 FAX: (765)4940558



