Name of Quantlet: Supervised-Randomization
Published in: Affordable Uplift: Supervised Randomization in Controlled Experiments (ICIS 2019)
Description: Supervised randomization integrates the targeting model into randomized controlled trials and A/B tests as a stochastic policy. By replacing random targeting with supervised randomization, we reduce the costs of running randomized experiments or allow continuous collection of randomized trial data. We show how to fully correct downstream analysis for the bias effect by the supervised treatment allocation with the true probabilites to receive treatment, which are logged at the time of targeting.
Keywords: CATE, RCT, AB-testing, Machine Learning
Author: Johannes Haupt, Daniel Jacob, Robin Gubela
See also:
Submitted: 26.07.2021
