Help using Base SAS procedures

Help on long exponential regression PROC nlin

Reply
N/A
Posts: 1

Help on long exponential regression PROC nlin

Hello all:

I have two questions related to nlin exponential regression


1.
proc nlin data=work.data1 best=5;
model Y = a+b*(1+exp(c*X)) ;
parameters a=-10 to 10 by .1 b=-10 to 10 by .1 c=-10 to 10 by .1;
output out=sr4 p=p4;
run;


How can i run such type of regression function, because it reaches the maximum iteration so it fails to converge. even if I put the parameters range very small, it still does not converge.


2.

Lets say i want to select the best-fit model in deferent way of exponential function.

e.g.

model1 Y = exp(f*X)
model2 Y= d*exp(d*X)
model3 Y= a+b*(1+exp(c*X))
.
.
.
.

is there any way to run all the models together so that the output contains the best-fit or performing function, or atleast can i set criteria?

Thanks in advance

SAS 9.1.3 Service Pack 4
XP_PRO platform
Ask a Question
Discussion stats
  • 0 replies
  • 85 views
  • 0 likes
  • 1 in conversation