Cox, M G; Harris, P M; Lazzari, A* (2004) The applicability of non-parametric methods of statistical analysis to metrology. NPL Report. CMSC 46/04

Text cmsc46.pdf Download (541kB)

Abstract

There are classes of problems in metrology for which little or no distributional knowledge is available concerning the values of the input quantities to a model of measurement. Consequently, non-parametric methods have value when applied to such problems. So-called ‘re-sampling’ methods are discussed and applied to two problems, one concerning environmental and the other dimensional measurement. These problems have very different character, and the re-sampling approaches used are selected accordingly. Further, consideration is given to an approach for the provision of a coverage region for the values of multivariate quantities that makes no assumption about the distribution for the values of these quantities. The approach is applied to a problem in electrical metrology.

Item Type:	Report/Guide (NPL Report)
NPL Report No.:	CMSC 46/04
Subjects:	Mathematics and Scientific Computing Mathematics and Scientific Computing > Modelling
Last Modified:	02 Feb 2018 13:16
URI:	http://eprintspublications.npl.co.uk/id/eprint/2919

Actions (login required)

View Item