Handling of skewed probability distributions for measurement uncertainties
|Coordinator||RISE Research Institutes of Sweden AB - Mätteknik Borås|
|Funding from Vinnova||SEK 860 000|
|Project duration||December 2018 - January 2021|
Purpose and goal
Aim: Develop a mathematical procedure for handling skewed distributions for measurement uncertainties. Improve estimations of of measurement uncertainties by considering skewness in the distributions. Increase the knowledge of skewness i measurement uncertainties. Objective: Develop a mathematical procedure for handling skewed distributions for measurement uncetainties. Increase knowledge of skewness in measurement uncertainties, and illustrate the relation between normal and log-normal distributions.
Expected results and effects
Improved evaluations of measurement uncertainties and their distributions at large measurement uncertainties that can occur in chemical analysis. Improved understanding of skewed uncertainty distributions.
Planned approach and implementation
A procedure that is intended to be used on data from control charts that describes intermediate precision that is easily accesable will be proposed. The following steps will be included: Development and optimization of an algorithm that make it possible to transform data to a normal distributed data set Characterization of the algorithm Application of the algorithm on small data sets (typically containing 100 data that corresponds to data in control charts) Discussion of issues, limitations and applicability