Can someone explain the "Mean -Log p " in the Fit Detail report in Logistic regression?
All I can find in the Help is this, and that a
"small number is better", but what is small?
and
"Mean -Log p: The average of -log(p), where p is the fitted probability associated with the event that occurred."
In my (really bad) model, I get :
Measure Training Validation Definition
Entropy RSquare 0,0333
Generalized RSquare 0,0789
Mean -Log p 1,0533
RASE 0,6449 0,6622
Mean Abs Dev 0,6352
Misclassification Rate 0,5502
