statsmodels.genmod.families.links.Logit.deriv¶

Logit.deriv(p)[source]¶

Derivative of the logit transform

Parameters:

p: array-like

Probabilities

Returns:

g’(p) : array

Value of the derivative of logit transform at p

Notes

g’(p) = 1 / (p * (1 - p))

Alias for Logit: logit = Logit()