It is based on model performance model
When to use kolmogorov smirnov test, GINI, AUC? Help check the below
3.1 Applying CART I used rpart package to solved this
3.2 Interpret the CART model output <pruning, remarks on pruning, plot the pruned tree> rpart package too
3.3 Applying Random Forests I used randomForest package for this
3.4 Interpret the RF model output <with remarks, making it meaningful for everybody>
4.1 Confusion matrix interpretation- I used caret package
**
> 4.2 Interpretation of other Model Performance Measures <KS, AUC, GINI>
this is where I am having a problem. I don't understand What the question is saying and the application too. Should they be applied on test, train or predicted values
I mean the 4.2, about KS, AUC, GINI #homework
**
Thank
It seems that you recently asked a question that originated from a class or workshop, and I wanted to make sure you were seeking help here in a way that maximizes the chances of getting good help (without violating our homework policies).
Having read your post, I'm afraid I see bullet points and not questions. Are you asking the forum, how to calculate statistics?, or how to interpret them ? or both ?
you can read about measuring performance with caret here
It shows how to get AUC.
Gini and AUC are just linear transforms of each other so
Gini = 2* AUC - 1
you can calculate Gini for any AUC with this formula.