I want to compare two roc curves generated from 2 probit models with different independent variables on same dependent binary variable.
I want to compute the t-stat and p-val of one roc curve being more significant than other.
Basically this is computing z in eqn 3 of this paper by Hanley Mcneil:
[login to view URL]
I want to know how they compute r. i.e how they generate Table 1
Is there any python fn which does this?