PerspectiveQuantifying uncertainty in qualitative analysis

Abstract

The feasibility of adopting a consistent approach to the expression of uncertainties relating to identification is discussed. It is argued that qualitative analysis can be viewed as a classification problem, that it is at least as important as quantitative analysis and that inferences drawn from qualitative tests should take relevant uncertainties into account. A brief review of systems of reasoning under uncertainty is presented, and it is concluded that Bayes' theorem provides the most suitable framework, providing for combination of separate items of evidence and implicitly allowing for both false positive and false negative probabilities in a single parameter. The chemical significance and practical evaluation of relevant probabilities are considered, and the applications and reporting of 'identification certainty' figures are discussed.

A tutorial on the validation of qualitative methods: From the univariate to the multivariate approach

“…This tutorial will explain how contingency tables can be used to estimate mandatory performance parameters. Detailed information on Bayesian decision theory [23] and statistical hypothesis tests [24,25] can be found elsewhere.…”

Validation, Uncertainty, and Quality Control of Qualitative Analysis of Tear Gas Weapons by Gas Chromatography-Mass Spectrometry

“…The evaluation of the examination uncertainty is based on the Bayes' theorem following theory and notation described by Ellison et al (1998).…”

“…Although the use of intrinsic standards in measurements (i.e., quantitative analysis) is not frequent, the examinations based on highly selective analytical instrumentation are often based on intrinsic and reproducible properties of standards. Different tools and metrics can be used to report examination uncertainty, such as Bayes' theorem, certainty factors, Dempster-Shafer theory, fuzzy sets, and so forth (Ellison, Gregory, and Hardcastle 1998). Nevertheless, the metrics of the Bayes' theorem are considered the most adequate for its generalized use as they are based on simple statistics, combine information of false positive and false negative rates in a single parameter, and can be used to estimate the improved quality of examinations supported on the combination of independent evidences of the event.…”

“…This work discusses the validation of the examination procedure, including the evaluation of the examination uncertainty, and the quality control of routine examinations. Examination uncertainty is determined using metrics of the Bayes' theorem (Ellison et al 1998).…”

Conditional/non‐conditional probabilities and information entropies assigned to high‐performance liquid chromatography with diode array and fluorescence detection

“…Bayes' formula provides a viable framework for assessing the uncertainty associated not only with qualitative analysis [15] but also with quantitative coupled processes of detection. …”