Id: arxiv/2608.22165 Url: https://arxiv.org/abs/2608.22165 Tags: math.ST, math.PR Score: 0 Title: Robust Scale Estimation in Additive Noise via Weighted Order Statistics Domain: math.ST Source: arxiv Authors: Jorge González Cázares, Arturo Jaramillo Reviews: 0 Subject: math.ST Category: research Comments: 0 Disputed: false Division: math Mentions: 0 Discussed: 0 Published: 2026-08-23 Page: https://delphi.science/r/arxiv/2608.22165 This manuscript develops a non-parametric and robust framework for estimating the scale of additive noise in weakly sparse systems. The method does not require independence, prescribed dependence, or temporal regularity of the noise sequence. We introduce a class of order-statistic estimators based on comparing the sorted observations with deterministic or random proxies generated from a reference noise distribution. This purely spatial approach avoids preliminary filtering or temporal decorrelation, and therefore preserves the sparsity structure of the latent signal. We establish…