# Une alternative robuste au maximum de vraisemblance: la   $\rho$-estimation

**Authors:** Yannick Baraud, Lucien Birg\'e

arXiv: 1703.01654 · 2017-07-04

## TL;DR

This paper introduces the $ho$-estimation method as a robust alternative to maximum likelihood estimation, highlighting its optimality and robustness properties through examples and connecting it to previous estimators.

## Contribution

It presents the $ho$-estimation framework, demonstrating its advantages over traditional estimators like MLE in terms of robustness and optimality across various statistical models.

## Key findings

- $ho$-estimators outperform MLE in robustness.
- The method offers optimal properties in diverse frameworks.
- Examples illustrate improved performance of $ho$-estimators.

## Abstract

This paper is based on our personal notes for the short course we gave on January 5, 2017 at Institut Henri Poincar\'e, after an invitation of the SFdS. Our purpose is to give an overview of the method of $\rho$-estimation and of the optimality and robustness properties of the estimators built according to this procedure. This method can be viewed as the sequel of a long series of researches which were devoted to the construction of estimators with good properties in various statistical frameworks. We shall emphasize the connection between the $\rho$-estimators and the previous ones, in particular the maximum likelihood estimator, and we shall show, via some typical examples, that the $\rho$-estimators perform better from various points of view.   ------   Cet article est fond\'e sur les notes du mini-cours que nous avons donn\'e le 5 janvier 2017 \`a l'Institut Henri Poincar\'e \`a l'occasion d'une journ\'ee organis\'ee par la SFdS et consacr\'ee \`a la Statistique Math\'ematique. Il vise \`a donner un aper\c{c}u de la m\'ethode de $\rho$-estimation ainsi que des propri\'et\'es d'optimalit\'e et de robustesse des estimateurs construits selon cette proc\'edure. Cette m\'ethode s'inscrit dans une longue lign\'ee de recherches dont l'objectif a \'et\'e de produire des estimateurs poss\'edant de bonnes propri\'et\'es pour un ensemble de cadres statistiques aussi vaste que possible. Nous mettrons en lumi\`ere les liens forts qui existent entre les $\rho$-estimateurs et ces pr\'ed\'ecesseurs, notamment les estimateurs du maximum de vraisemblance, mais montrerons \'egalement, au travers d'exemples choisis, que les $\rho$-estimateurs les surpassent sur bien des aspects.

## Full text

_Full body text omitted from this summary view._ Fetch the complete paper as Markdown: https://tomesphere.com/paper/1703.01654/full.md

## References

14 references — full list in the complete paper: https://tomesphere.com/paper/1703.01654/full.md

---
Source: https://tomesphere.com/paper/1703.01654