Precise Measurand Value Estimating by Interval Fusion with Preference Aggregation: Heteroscedasticity Case

Sergey V Muravyov, Liudmila I Khudonogova, Dai Minh Ho

Abstract

It is considered the problem of determination of a reference value for heteroscedastic interval data. For this aim, the newly proposed by the authors interval fusion with preference aggregation (IF&PA) procedure is used. The procedure, modified to improve the accuracy of the fusion result, is presented and applied to process the heteroscedastic data of a real experiment. The experiment consisted in determination of the reference value of DC voltage based on the readings of five different models of multimeters. For comparison, the same data were processed by the method of weighted mean. For two methods, an absolute deviations of the obtained reference values from the nominal value (which is a high-precision calibrator output) and reference value uncertainties were estimated. It is shown that the IF&PA procedure allows to obtain a reference value very close to nominal value and with considerably reduced uncertainty in comparison with traditional method based on weighted mean calculation.

Keywords

Interval fusion, Preference aggregation, Heteroscedasticity, Consensus estimate, Ranking

Download

IMEKO-TC10-2020-029.pdf

DOI

-

IMEKO TC

TC10 - Measurement for Diagnostics, Optimization and Control

Back to the proceedings