I have a response threshold that I have calculated using a sigmoidal curve fitted to my data set. What I want to understand is why this is a good method for finding a threshold over maybe setting a rule for when you detect a response? What makes a sigmoidal model good for generating threshold values? I have often seen it in the context of neural networks but am struggling to find it relating to biological thresholds.
Any insight would be appreciated
There are individual differences in the organism, it is difficult to find a generating threshold values.
A sigmoidal curve is mathematically easy to handle because it has a non-negative derivative at each point. A threshold value can be calculated e.g. by setting the second derivative to zero. If no other model is availabe a sigmoidal curve is often used.
https://en.wikipedia.org/wiki/Sigmoid_function
- Badges
- Science topic