[Date Index] [Thread Index] [Author Index]

Re: Nonlinear Fit Problem

• To: mathgroup at smc.vnet.net
• Subject: [mg7399] Re: Nonlinear Fit Problem
• From: Tom Burton <tburton at cts.com>
• Date: Fri, 30 May 1997 01:20:12 -0400 (EDT)
• Sender: owner-wri-mathgroup at wolfram.com

>> The difference is within the default precision for the method. No
>>problems, as far as I can tell, with either of our solutions. 
> 
>> Tom Burton
>
>
>Tom, thank you for your response.  I think an error still exists with the
>fit routine.  Plot datafit (Out[83]) from 0 to 5 and on the same plot
>display the data that is being fit (data7).  You will notice that the fit
>doesn't approximate the data.  It should, no matter how many constants are
>added to the fitting function. 
>
>David Junkin
>
>
Good grief--you're right! I was confusing the model output with parameters
output. I tried your problem also with NonlinearRegress. The value returned
for "a" is unaffected by the constant in the model. This indeed appears to
be an error.

Sorry for my earlier post.

Tom Burton