🚀 go-pugleaf

SAS and SYSTAT both do general non-linear fitting. Both give default
criteria for finding the "Best Fit" and both allow you to enter your own
objective function. There are, for example, maximum likelihood methods,
minimum chisquare methods, and non-linear minimum squared error methods. 

Unlike simple linear methods, the fit you get out of one program, (or
even from one first guess) won't necessarily look like the fit you get
under other circumstances. It's also easy to accidentally specify a
model that cannot be identified. 

If the software doesn't give you what you're expecting, do some reading
or consult someone who has. 

Luck and Regards,
-LArry C.