Dear Maarten and Dave,
Thank you very much for the comments and the detailed mail and code. I knew that one could estimate the knots using nonlinear regression, but I wasn’t sure how. The attached code helps me a lot in that respect. As you mentioned in you mail, the inflection point can be found by equating the second derivative to zero. But this can’t be done with the exponential function that I have.
Estimating the derivative of the lowess curve is an excellent idea. I’m definitely interested in the paper and the code. Please mail them to me when you get time.
Thank You
Leny Mathew
begin:vcard
n:Mathew;Leny
fn:Leny Mathew
tel;work:215 762 2069
org:Drexel University College of Medicine;OB/GYN control
version:2.1
email;internet:[email protected]
title:Data Analyst
end:vcard
