Welcome to kuplift’s documentation!

Description

It’s a User Parameter-free Bayesian Framework for Uplift Modeling.

How to use kuplift ?

A note on the terminology used in the source code

The source code uses short names such as i, j or t. Here are their definitions:

  • i: part (interval for a numerical variable or value group for a categorical variable);

  • j: target (outcome);

  • t: treatment;

  • g: group of treatments;

  • N: number of observations (frequency);

  • P: probability.

The source code also refers to tables, represented by pandas.DataFrame and named using these short names. For example: N_ijt, P_ijg, Uplift_ig. Explanation of the example names:

  • N_ijt: Each value in the table is a number of observations. One DataFrame row contains the values for one part (i). One DataFrame column contains the values for one target-treatment pair (jt).

  • P_ijg: Each value in the table is a probability. One DataFrame row contains the values for one part (i). One DataFrame column contains the values for one target-treatmentgroup pair (jg).

  • Uplift_ig: Each value in the table is an uplift. One DataFrame row contains the values for one part (i). One DataFrame column contains the values for one treatment group (g).