I am confused with the Pseudo $R^2$ computation of the negative binomial regression model. For the logistic regression, we can compute the Pseudo $R^2$ as
pR = 1 - fit$deviance /fit$null.deviance pR
I believe that the interpretation of this quantity $pR$ will be the amount of deviance explained by the regression model.
Can we use this for the negative binomial regression model? Thank you so much for your time!