B0598
Title: A computationally efficient framework for robust estimation
Authors: Yuming Zhang - University of Geneva (Switzerland) [presenting]
Samuel Orso - University of Geneva (Switzerland)
Maria-Pia Victoria-Feser - University of Geneva (Switzerland)
Stephane Guerrier - University of Geneva (Switzerland)
Abstract: Constructing estimators that are robust to data contamination is non-trivial. Indeed, to be consistent, these estimators typically rely on a non-negligible correction term with no closed-form expression. Numerical approximation to this term can introduce finite sample bias, especially when the number of parameters $p$ is relatively large compared to the sample size $n$. To address these challenges, a simulation-based bias correction framework is proposed, which allows easy construction of robust estimators with reduced finite sample bias. The key advantage of the proposed framework is that it bypasses the computation of the correction term in the standard procedure. The resulting estimators also enjoy consistency and asymptotic normality and can be obtained computationally efficiently even when $p$ is relatively large compared to $n$. The advantages of the method are highlighted in different simulation studies, such as logistic regression and negative binomial regression models. It is also observed empirically that the estimators are actually comparable, in terms of finite sample mean squared error, to classical maximum likelihood estimators under no data contamination.
