Bottom-up and top-down uncertainty quantification for measurements

被引：9
|
作者：
Burr, T. [1 ]
Croft, S. [2 ]
Favalli, A. [3 ]
Krieger, T.
Weaver, B. [1 ]
机构：
[1] Los Alamos Natl Lab, Stat Sci Grp, Los Alamos, NM USA
[2] Oak Ridge Natl Lab, Nucl Secur, Oak Ridge, TN USA
[3] Los Alamos Natl Lab, Nucl Sci & Technol, Los Alamos, NM USA
来源：
CHEMOMETRICS AND INTELLIGENT LABORATORY SYSTEMS | 2021年 / 211卷
关键词：
Approximate Bayesian computation (ABC); Bottom-up and top-down UQ (UQ); Data-driven choices; Guide to expression of uncertainty in; measurement; Item-specific bias; t-distribution; ERROR VARIANCE;
D O I：
10.1016/j.chemolab.2020.104224
中图分类号：
TP [自动化技术、计算机技术];
学科分类号：
0812 ;
摘要：
Several recent papers address improved uncertainty quantification (UQ) for measurements used in nuclear safeguards. This paper reviews progress and presents new results for bottom-up (first principles) and top-down (empirical) UQ for safeguards, where the main quantitative measure of uncertainty is the total measurement error standard deviation (SD), which includes both random and systematic error components. The five main UQ topics addressed here include: (1) impact of making data-driven choices in SD estimation; (2) use of approximate Bayesian computation (ABC) for both bottom-up and top-down UQ; (3) computational calibration; (4) revisions to the guide to the expression of uncertainty in measurement (GUM), and (5) critique of a recently-suggested ?Unified Theory of Measurement Errors and Uncertainties.?
引用
收藏
页数：14
相关论文
共 50 条
12345