Metrological Certification of the Algorithm for Estimating Inconsistent Data of Key Comparisons of National Standards

• Published in: Measurement techniques Vol. 65; no. 7; pp. 508 - 514
• Main Authors: Burmistrova, N. A., Viktorov, I. V., Chunovkina, A. G.
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.10.2022; Springer; Springer Nature B.V
• Subjects: Algorithms; Analysis; Analytical Chemistry; Certification; Characterization and Evaluation of Materials; Data processing; Estimates; General Questions of Metrology and Measurement Technology; Measurement Science and Instrumentation; Physical Chemistry; Physics; Physics and Astronomy; Statistical models; Uncertainty; measurement uncertainty; dark uncertainty; metrological compatibility; REM algorithm; inconsistent data
• ISSN: 0543-1972; 1573-8906
• DOI: 10.1007/s11018-023-02111-1

An important stage in the preparation of a report on key comparisons of national standards is considered — data processing using various algorithms. We investigate properties of estimates of the algorithm for processing inconsistent data of comparisons based on the random effects model (REM). This model is used to describe the so-called dark uncertainty (undetected uncertainty). When evaluating the properties of the algorithm, the methodology of metrological certification of algorithms was applied, which is based on the study of the algorithm on typical data models that correspond to practical situations. Using the method of statistical modeling, the following characteristics of the dark uncertainty estimates were obtained: the conditional distribution density of the estimate, the standard deviation, and the bias. Based on the research results, practical recommendations for the use of the REM algorithm are formulated.