This is a tricky one, but I think Option B is the way to go. The lift chart is all about improving the model's performance, so it's about the number of events, not accuracy.
I'm leaning towards Option D. Selecting the observations with a 10% response probability should result in 3.14 times greater accuracy compared to a random draw.
Option B seems correct. The lift chart is showing the lift at a certain depth, which means the selected observations should have 3.14 times more events than a random draw of the same size.
Olive
5 hours agoMarlon
4 days agoLatrice
12 days agoErasmo
21 days agoLoreta
22 days agoLenna
25 days ago