Date:         Sat, 12 Oct 1996 05:11:57 GMT
Reply-To:     Michael Benson <mbenson@IAC.NET>
Sender:       "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From:         Michael Benson <mbenson@IAC.NET>
Organization: Internet Access Cincinnati 513-887-8877
Subject:      Re: regression w/ nonlinear constraints

You could replace, say, A in your model with (1 - B1) * B2 / B3 and then
use  PROC NLIN  to get all the B's and then calculate A afterwards.
 
In article <531b71$kh2@geraldo.cc.utexas.edu>, chris bratcher
<bratcher@jeeves.la.utexas.edu> wrote:
 
>The model is:
>
>Yi = A  + B1*X1  + B2*X2 + B3*X3 + . . .
>
[with]
>A/(1 - B1) = B2/B3
 
Michael
--
Michael Benson    mbenson@iac.net    Cincinnati, OH
--