Figures of Merit in Multiway Calibration

Abstract

As previously presented in this book, measuring and processing multiway data provides analytical chemists with a number of advantages, such as (1) improved sensitivity, derived from noise averaging multiple measurements of redundant data, (2) increased selectivity, because each new data mode provides an additional degree of partial selectivity, and (3) modeling the analyte contribution and its quantitative determination in the presence of unknown interferences, absent in calibration samples (second-order advantage) [1]. Regarding items 1 and 2, a question which immediately emerges is how figures of merit like sensitivity, selectivity, and even the limit of detection (LOD) should be estimated when dealing with multivariate and multiway data? As analytical chemistry is the science of chemical measurements, finding a reliable way to judge them properly is not a minor issue, and this is the reason why this chapter is focused on trying to give a response to this question.