A comparative study of counterfactual estimators
Machine Learning
2019-01-30 v3 Machine Learning
Abstract
We provide a comparative study of several widely used off-policy estimators (Empirical Average, Basic Importance Sampling and Normalized Importance Sampling), detailing the different regimes where they are individually suboptimal. We then exhibit properties optimal estimators should possess. In the case where examples have been gathered using multiple policies, we show that fused estimators dominate basic ones but can still be improved.
Cite
@article{arxiv.1704.00773,
title = {A comparative study of counterfactual estimators},
author = {Thomas Nedelec and Nicolas Le Roux and Vianney Perchet},
journal= {arXiv preprint arXiv:1704.00773},
year = {2019}
}