Extreme residuals in regression model. Minimax approach

TL;DR

This paper develops limit theorems for the behavior of extreme residuals in linear regression models when using minimax estimation of parameters, providing insights into the distribution of worst-case errors.

Contribution

It introduces new limit theorems specifically for extreme residuals under minimax estimation in linear regression, advancing understanding of worst-case error behavior.

Findings

Limit theorems for extreme residuals established

Characterization of residual distribution under minimax estimation

Insights into worst-case error behavior in regression models

Abstract

We obtain limit theorems for extreme residuals in linear regression model in the case of minimax estimation of parameters.