0
$\begingroup$

I'm looking for a general function that can fit experimental data that has a profile/trend similar to the logit function $f(x) = \log x/(1-x)$, but where all f(x) values are between 0 and 1.

Is there a generalised version of the logit function, or any other transformed sigmoidal function, that I can scale and fit my data below?

f(0.005) = 0.005

f(0.04) = 0.06

f(0.1) = 0.22

f(0.2) = 0.37

f(0.3) = 0.47

f(0.4) = 0.48

f(0.5) = 0.53

f(0.6) = 0.55

f(0.7) = 0.56

f(0.8) = 0.58

f(0.9) = 0.66

f(0.96) = 0.79

f(0.999) = 1.00

I'm also attaching an image of the plot. Thanks!

curve_fit_problem

$\endgroup$
  • $\begingroup$ what about the probit function? $\endgroup$ – Fede Poncio Jul 14 '17 at 22:56
0
$\begingroup$

What about this one? $$f(x)=4(x-0.5)^3+0.5$$

There is no logarithm or exponential, but the shape of the curve is very similar.

$\endgroup$
0
$\begingroup$

It's quite a good fit to $$ x + .684 x (1-x)+8.98 x^2 (1-x)-48.6 x^3 (1-x)+79.3 x^4 (1-x)-45.7 x^5 (1-x) $$ obtained by linear least squares.

enter image description here

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.