1

I've got the following:

x <- c(549.9 , 400 , 325.9 , 250.2 , 200.1 , 150 , 120 , 79.92 , 39.96)

y <- c(5.748 , 3.50152 , 2.70204 , 2.54265 , 2.52241 , 1.35861 , 1.377444444 ,  1.279055556 , 1.05501)

nls(y ~ A + B*(x^C))

I haven't managed to make it work so far. Am I missing something?

Also, how can I make my initial guesses for the parameter A, B and C?

bouteillebleu
  • 2,456
  • 23
  • 32
Gilmar Neves
  • 37
  • 10
  • It's not the final aswer to this but my post on CrossValidated had some comments quite useful to this kind of problem. https://stats.stackexchange.com/questions/328859/comparing-fitting-methods-for-the-same-model-bland-altman-and-mann-whitney – Gilmar Neves Jul 13 '20 at 11:58

1 Answers1

0

The most direct answer is simply the documentation. I missed some parameters. https://www.rdocumentation.org/packages/stats/versions/3.6.2/topics/nls

For those looking for a similar thing, I have got quite useful information when I started a similar discussion o CrossValidated.

Here are the links in which you can get to know a bit about the subject:

https://stats.stackexchange.com/questions/328859/comparing-fitting-methods-for-the-same-model-bland-altman-and-mann-whitney

https://stats.stackexchange.com/questions/357977/does-it-make-sense-having-scale-and-log-transformed-variables-to-select-models

Gilmar Neves
  • 37
  • 10