IMPROVED BOUNDS FOR SMALL-SAMPLE ESTIMATION

被引：14
|
作者：
Gratton, Serge [1 ]
Titley-Peloquin, David [2 ]
机构：
[1] Univ Toulouse, INPT IRIT ENSEEIHT, F-31071 Toulouse, France
[2] McGill Univ, Dept Bioresource Engn, Montreal, PQ H3A O39, Canada
来源：
SIAM JOURNAL ON MATRIX ANALYSIS AND APPLICATIONS | 2018年 / 39卷 / 02期
关键词：
statistical estimation; small-sample estimation; norm estimation; condition number estimation; STATISTICAL CONDITION ESTIMATION; GAUSSIAN-PROCESSES; CONDITION NUMBERS;
D O I：
10.1137/17M1137541
中图分类号：
O29 [应用数学];
学科分类号：
070104 ;
摘要：
We derive improved error bounds for small-sample statistical estimation of the matrix Frobenius norm. The bounds rigorously establish that small-sample estimators provide reliable order-of-magnitude estimates of norms and condition numbers, for matrices of arbitrary rank, even when very few random samples are used.
引用
收藏
页码：922 / 931
页数：10
相关论文
共 50 条
12345