A COMPUTATIONAL METHOD IN THE CONSTRUCTION OF A PROBABILITY REGION OF CONFIDENCE

G. Iuculano, A. Zanobini, G. Pellegrini

Abstract

The construction of a region of confidence or "confidence belt", that is the expression of an expanded uncertainty with a given coverage probability, for measurements performed on the same measurand by different processes, is a critical task in the perspective to establish a measurement traceability on a global basis. An alternative procedure to the analytical one suggested in the Guide to the Expression of Uncertainty in Measurement, strictly related to the validity of the Central Limit Theorem, is desirable, in order to cover a larger range of situations not included in the normal distribution cases. In this work statistical approach is proposed based on the use of Monte Carlo approximation technique and on the bootstrap resampling iteration. An experimental model is examined to check the validity of the proposed method.

Keywords

Confidence Belt, Monte Carlo-Bootstrap Approximation

Download

IMEKO-WC-2000-TC8-P224.pdf

DOI

-

IMEKO TC

TC8 - Traceability in Metrology

Back to the proceedings