This is standard stuff covered in the help and
the manuals or learnable by simple experiments.
But in short
_b[]
is Stata standard notation for a parameter (coefficient) in
a model. Usually a varname belongs in the [], with
the important exception _cons. But with non-linear
models there is typically not a one-to-one correspondence
between parameters and variables -- else they would be
linear -- and so the notation /parametername is the
alternative adopted. Or so I surmise.
Try
. sysuse auto
. gen gpm = 1/mpg
. nl (gpm = {a}*weight)
or any other example. (I know it's linear).
Nick
n.j.cox@durham.ac.uk
Kyle Hood
> what does the notation _b[/rho] do? Why _b[]? why /?
Oleksandr Shepotylo wrote:
>
> > Thank you,
> >
> > I need this rho to transform my exogenous variables X(NxK)
> by mulitplying
> > them by the following matrix:
> >
> > X*=rho*W*X, where W is (NxN)
> >
> > so I planned to calcualte rho*X and use mata to calculate W*(rho*X).
> >
> > But you are right, I can save a scalar rho and use it in
> mata expressions
> > when I need it.
Nick Cox wrote:
> >> This sounds like a bad idea. The parameter can be saved as
> a scalar or a
> >> local macro, but using a variable for
> >> the same number again and again is usually unnecessary.
> >> Still, there might be a good idea, or you might insist, so
> if so try
> >> gen rho = _b[/rho]
> >> but consider also
> >>
> >> scalar rho = _b[/rho]
Oleksandr Shepotylo
> >>> After non linear estimaition which estimates parameters
> rho and sigma of
> >>> the following form
> >>>
> >>> nl (z=h1*{rho}+h2*{rho}^2+h3*{sigma2}), in(rho -0.5
> sigma2 1) iterate(200)
> >>>
> >>> I want to generate a variable rho that takes the value of
> the estimated
> >>> parameter {rho}
> >>>
> >>> gen rho=???
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/