A/B testing with Random Forest

General non-parametric A/B test based on the Random Forest using the R-package hypoRF

Figure 1: figure from the original paper (by authors) “On the Use of Random Forest for Two-Sample Testing” showing the p-value of the test and the variable importance (left) as well as the projection of the data on the two discriminative components (right).

Contributors: Loris Michel, Jeffrey Näf

The importance of A/B testing in business decisions is not a new topic. We can point the reader to this story to get a good recap about the major steps and challenges underlying A/B testing.