scikit-uplift
scikit-uplift copied to clipboard
Generating a random uplift dataset
💡 Feature request
We need a function to generate a random binary uplift task make_uplift_classification.
It is contains three columns: target, treatment and uplift. Target and treatment columns are binary.
We need some requirements for the result. They are:
- reproducibility
- uplift continuality (there should not be many identical values for a large n = number of observations)
- clear parameters to change the uplift curve closer to the ideal and closer to the random one
- a random uplift must be non-negative, you need to be able to change it.