Skip to content
Prev 268104 / 398502 Next

nls, how to determine function?

Hi Petr, thanks for your help on this. I will most definitely get this
book as it appears to be a good one.
I am happy with how nls() and the self starting function appears to be
fitting the data set.

What I am wondering about is how to write out this function outside of R.
For example, if I sat down with a paper and pencil and wrote this
function out in terms of f(x) = ?. In this example
(http://www.graphpad.com/curvefit/id203.htm) there is a formula
contains for what they are calling the Boltzmann Sigmoid function. In
order to have an answer, is this something I need to generate on my
own? Or, is there a common form that is being applied by the SSlogis
and the SSfpl functions? In Pinheiro and Bates (pg 518) there is a
model  f(x) = A + B/1+ exp[(scal-x)/expL]. Is this the form of the
equation?

Or, is there a way to extract the local min and max values from nls ()
results? Do I have to break apart the individual components of the
regression into piecewise functions?

Thanks again for your help.

Katrina
On Tue, Aug 9, 2011 at 12:38 AM, Petr PIKAL <petr.pikal at precheza.cz> wrote: