PRIOR INFORMATION IN PRODUCT-RATIO MEASUREMENTS

V. Tuninsky, W. Wöger

Abstract

In this paper we consider direct repeated measurements of a quantity <i>R</i> that is defined by the product <i>P</i>₁^<i>n</i>₁ &middot; <i>P</i>_<i>m</i>^{<i>n</i>_<i>m</i>} (<i>n</i>₁, ... ,<i>n</i>_<i>m</i> = &plusmn;1 ) of <i>m</i> other quantities. Bayes theorem is applied to obtain posterior estimates of <i>R</i> and <i>P</i><i>&alpha;</i> (<i>&alpha;</i> = 1, ..., <i>n</i>) based on the measurements of <i>R</i> and prior information (estimates, associated uncertainties) about <i>P</i><i>&alpha;</i>. The paper extends a previous result applicable to a calibration chain with only two elements to chains with arbitrary number of elements.

Keywords

evaluation of measurement results, traceability in metrology

Download

IMEKO-WC-2000-EVL-P546.pdf

DOI

-

Back to the proceedings