Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down at the end of May, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: non linear equation difficulties with nl function evaluation program


From   Austin Nichols <austinnichols@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: non linear equation difficulties with nl function evaluation program
Date   Wed, 25 Aug 2010 10:45:25 -0400

Carolina Lopez <carolina.univ@gmail.com>:
Also consider whether there are theoretical limits on parameters that
-nl- could not know about, and which could lead it into infeasible
areas of the parameter space; for example, if k is always positive,
you might prefer to specify lnk as a parameter and then put exp(`lnk')
in your formulas.

On Wed, Aug 25, 2010 at 7:26 AM, Carolina Lopez <carolina.univ@gmail.com> wrote:
> Thanks both for your inputs.
>
> I´ll review the model again to see how I can simplify and fit it,
> focusing on k which is my key parameter.
>
> Carolina
>
> 2010/8/25 Nick Cox <n.j.cox@durham.ac.uk>:
>> I agree with Maarten's stance here.
>>
>> Has anyone actually fitted such a model with a success in your area? There are plenty of fields in which complicated models are copied from paper to paper and book to book, yet no-one has ever properly tested or fitted them.
>>
>> There is plenty of cautionary literature, e.g.
>>
>> http://demonstrations.wolfram.com/FittingAnElephant/
>>
>> Wei, J. 1975. Least square fitting of an elephant. CHEMTECH 5: 128–129.
>>
>> Nick
>> n.j.cox@durham.ac.uk
>>
>> Maarten buis
>>
>> --- On Wed, 25/8/10, Carolina Lopez wrote:
>>> I did start with a simple equation, which converged and
>>> gave significant parameters. But when I extended it to this
>>> complete version, I started finding issues.
>>>
>>> The thing is as alpha, k and n are in all my terms, I
>>> cannot do (or do not know how) independent programs that
>>> then get built up on one equation to estimate.
>>
>> Removing paramters is one way of simplifying an equation,
>> you can also remove other parts, or change the equation in
>> other ways. I cannot explain it more clearly than that you
>> need to look at the equation to find a suitable
>> simplification. Sounds simple, but I know it is hard and
>> frustrating work.
>>
>> The plus of this is, is that once you are done, you'll not
>> only be able to estimate your model, you will also have a
>> clearer understanding of it. It is not much of comfort now,
>> but it may be one in the future.
>>
>>

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index