Hi all,

I'm wondering if it's possible to run nonlinear regression,
transformed into linear space, with a dummy variable.  For instance,
you can run

y = A^b1 

as

ln y = b1*ln(a),

but what if I was interested in something like

ln y = b1*ln(a) + b2*D, 

where D is 0 or 1. Is this possible?  

I get the nonlinear, back-transformed equation looking like

y = A^b1 * e^(b2*D).

Does this last equation make sense??  I think it does, but that may be
because of my last two Budweisers.

Thanks a lot,
Jason
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to