http://www.bing.com:80/search?q=Compte+Admin搜索结果http://www.bing.com:80/s/a/rsslogo.gifhttp://www.bing.com:80/search?q=Compte+Admin版权所有 © 2026 Microsoft。保留所有权利。不得以任何方式或出于任何目的使用、复制或传输这些 XML 结果，除非出于个人的非商业用途在 RSS 聚合器中呈现必应结果。对这些结果的任何其他使用都需要获得 Microsoft Corporation 的明确书面许可。一经访问此网页或以任何方式使用这些结果，即表示您同意受上述限制的约束。https://arxiv.org/abs/1702.00482We study the problem of estimating the mean of a random vector X given a sample of N independent, identically distributed points. We introduce a new estimator that achieves a purely sub-Gaussian performance under the only condition that the second moment of X exists. The estimator is based on a novel concept of a multivariate median.周六, 24 10月 2026 08:23:00 GMThttps://arxiv.org/pdf/1702.00482v1The proof of Theorem 2 is based on the idea of “median-of-means tour-naments” which was introduced by Lugosi and Mendelson [15] is the context of regression function estimation.周日, 10 5月 2026 01:33:00 GMThttps://projecteuclid.org/journals/annals-of-statistics/volume-47/issue-2/Sub-Gaussian-estimators-of-the-mean-of-a-random-vector/10.1214/17-AOS1639.pdfWe study the problem of estimating the mean of a random vector \( X \) given a sample of \( N \) independent, identically distributed points. We introduce a new estimator that achieves a purely sub-Gaussian performance under the only condition that the second moment of \( X \) exists.周六, 10 10月 2026 15:35:00 GMThttps://projecteuclid.org/journals/annals-of-statistics/volume-47/issue-2/Sub-Gaussian-estimators-of-the-mean-of-a-random-vector/10.1214/17-AOS1639.fullWe introduce a new estimator that achieves a purely sub-Gaussian performance under the only condition that the second moment of 𝑋 exists. The estimator is based on a novel concept of a multivariate median. Gábor Lugosi. Shahar Mendelson. "Sub-Gaussian estimators of the mean of a random vector."周一, 26 10月 2026 00:00:00 GMThttps://bse.eu/research/publications/sub-gaussian-estimators-mean-random-vectorWe introduce a new estimator that achieves a purely sub-Gaussian performance under the only condition that the second moment of X exists. The estimator is based on a novel concept of a multivariate median.周一, 05 10月 2026 05:54:00 GMThttps://zhuanlan.zhihu.com/p/662713569引入 sub-Gaussian分布 的动机，主要是考虑Gaussian分布的强尾衰减： 定义. 满足以下性质的 随机变量 X ∈ R. 根据 Markov's Lemma 容易推导以下定理. 定理. 若 X ∼ subG. 证明. 利用一个常用的 Chernoff bound 技巧： 由于 s 是我们任意取的待定值，取适当的 s 得证. 性质. 与Hoeffding' inequality的联动. 定理. 记 X 为随机变量且 \mathbb {E} (X)=0 ，以及 X \in [a, b] 几乎处处成立.周三, 21 10月 2026 01:53:00 GMThttps://luc.devroye.org/Devroye+Lerasle+Lugosi+Oliveira-SubgaussianEstimators-2016.pdfRecently, there has been a surge of interest in sub-Gaussian estimators, their generalizations to multi-variate settings, and their applications in a variety of statistical learning problems where heavy-tailed distributions may be present; see, for example, Catoni [5], Hsu and Sabato [9], Brownlees, Joly and Lugosi [3], Lerasle and Oliveira [14 ...周五, 23 10月 2026 13:54:00 GMThttps://tselilschramm.org/sos-paradigm/notes21/04-heavy-tailed-means.pdfLugosi and Mendelson introduce a new notion of high-dimensional median which tries to minimize the maximum discrepancy between its projection on u and the u-directional median among all u.周三, 19 8月 2026 18:53:00 GMThttps://www.jstor.org/stable/44245766In this paper, we prove that, for most (but not all) classes V C V2 we consider, there do exist estimators that achieve (5) for all large n , with Smin ^ e~cvn and a value of L that does not depend on S or n.周五, 09 10月 2026 17:24:00 GMThttps://www.researchgate.net/publication/313247464_Sub-Gaussian_estimators_of_the_mean_of_a_random_vectorA line of work (Lugosi & Mendelson, 2017; Hopkins, 2018;Cherapanamjeri et al., 2019) has shown how to achieve the subgaussian rate, ignoring constants, in polynomial time.周五, 21 4月 2023 13:48:00 GMT