Abstract
We study the problem of estimating the common mean μ of n independent symmetric random variables with different and unknown standard deviations . We show that, under some mild regularity assumptions on the distribution, there is an adaptive estimator such that it is invariant to permutations of the elements of the sample and satisfies that, up to logarithmic factors, with high probability,
where the index satisfies .
Nous étudions le problème de l’estimation de la moyenne commune μ de n variables aléatoires symétriques indépendantes avec des écarts types différents et inconnus . Nous montrons que, sous faibles hypothèses de régularité sur la distribution, il existe un estimateur adaptatif invariant par rapport aux permutations des éléments de l’échantillon qui satisfait à facteurs logarithmiques près et avec une grande probabilité
où l’indice satisfait .
Funding Statement
Luc Devroye was supported by NSERC Discovery Grants and by an FRQNT Team Research Grant. Gábor Lugosi was supported by the Spanish Ministry of Economy and Competitiveness, Grant PGC2018-101643-B-I00 and FEDER, EU, and by “Google Focused Award Algorithms and Learning for AI”. Nikita Zhivotovskiy is funded in part by ETH Foundations of Data Science (ETH-FDS).
Acknowledgments
The authors would like to thank the anonymous referees and an Associate Editor for their comments that improved the presentation of the paper.
Citation
Luc Devroye. Silvio Lattanzi. Gábor Lugosi. Nikita Zhivotovskiy. "On mean estimation for heteroscedastic random variables." Ann. Inst. H. Poincaré Probab. Statist. 59 (1) 1 - 20, February 2023. https://doi.org/10.1214/21-AIHP1239
Information