Qini Curve plot about scikit-uplift HOT 1 OPEN

Hello! Thanks for a question.

Sometimes, for test data, we don't know the number of people we need to communicate with. If we know the exact amount of the people we want to communicate with, we would use other metrics, for example uplift@k.

In the case of a Qini curve along the x-axis, all objects from the test group are really located. However, the axis is called "Number targeted" so because we can look at the Qini value at all possible communication group values.

You can read more about the Qini curve in the original article: Radcliffe, N.J. (2007). Using control groups to target on predicted lift: Building and assessing uplift model. Direct Market J Direct Market Assoc Anal Council, 1:14–21, 2007.

from scikit-uplift.