Re: st: Multiplicative Functions in Stata

[Nick Cox <njcoxstata@gmail.com>]

You don't say why you want to "input" this to Stata. My best guess is
that you want to fit this model, for which you need -nl-.

Looking at

(a+X^b+Y^c+Z^d)[1+exp(u-sum(siTi)-nM)]^(-1)

the reciprocation is easier to write as a division. Stata has its own
meanings for square brackets, which should become parentheses

(a+X^b+Y^c+Z^d)/(1+exp(u-sum(siTi)-nM))

Multiplications must all be explicit

(a+X^b+Y^c+Z^d)/(1+exp(u-sum(siTi)-n*M))

and the sum will need to be spelled out (e.g.)

(a + X^b  + Y^c + Z^d)/(1 + exp(u - s1 * T1 - s2 * T2 - s3 *T3 - n*M))

I will stop there. The -nl- manual entry says more, especially about
the brace notation to indicate parameters.

-nl- often needs good initial estimates of the parameters to converge
at all, as is characteristic of the genre.
Nick
njcoxstata@gmail.com


On 30 November 2013 02:57, Justin Hustrulid <jhustrulid@gmail.com> wrote:
> Hello,
>
> I have a damage abatement model I've seen in an article that I'm trying to
> replicate, but I'm not quite sure how to input it into Stata.  I'm using
> Stata 12 SE for Windows.
>
> Here's the function:
>
> Y=F(Z)G(X)
>
> Y=(a+X^b+Y^c+Z^d)[1+exp(u-sum(siTi)-nM)]^(-1)
>
>
>
> I hope its clear, as I had to retype it since the original was in html.
> X,Y, Z, T, and M are independent variables, while the lower case letters
> are parameters to be estimated.  i is an index, but I can't get it to be
> subscript without statalist bouncing the email back to me.  The first part
> is a straightforward log-log Cobb-Douglass function, but
> I'm not sure how to do the part in brackets or how to multiply them
> together to get Y.  I'm also unsure of whether both parts are done at the
> same time or in separate regressions.  The article says the G(X) part is
> logistic.  I'm fairly new to Stata and econometrics in general so any help
> in making this regression work would be greatly appreciated.  I hope I can
> learn a lot from statalist.
>
> Thanks,
> Justin
> *
> *   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/
*
*   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/