I am using a logistic function to model observed response values (growth) as function of an input value:
$$\text{growth} = \displaystyle{\frac{A}{1+e^{\frac{k-\text{input}}{s}}}}$$
This functional form has lower asymptote of growth=$0$, but at input=$0$ it can assume a positive growth value.
I would like to use a similar sigmoid function with the constraint of growth=$0$ at input=$0$.
How can I modify the function?
I found this but it did not solve the problem: Logistic function passing through two points AND the origin?
$$ s(x) = \frac{1}{1+\exp(-x)} $$
then:
$$ \tilde{s}(x) = s(x) - \frac{1}{2} $$
goes through the origin at $x=0$. $\tilde{s}$ is a translated sigmoid function.
– pitchounet Sep 19 '17 at 09:15