Multivariate analytical sensitivity in the determination of selenium, copper, lead and cadmium by stripping voltammetry when using soft calibration

Uncertainty estimation and figures of merit for multivariate calibration (IUPAC Technical Report)

“…-Analytical sensitivity. It is defined for first-order [145][146][147] and for second-order bilinear calibration [148], using a direct analogy with the univariate parameter. -Limit of detection.…”

Evaluation of analytical figures of merit when multivariate calibration is employed in atomic spectrometry

“…As the error in the reference values was assumed negligible because the preparation of the calibration emulsions is more accurate than the measurement and prediction steps, s y/x can be considered as a good approximation to the repeatability of the overall methodology (measurement plus multivariate modelling). 21 As mentioned above (section on Validation of the calibration model), it was 0.23 mg Cu L À1 which was satisfactory because it represents a 4.6% relative standard deviation (RSD) for a nominal concentration of 5 mg Cu L À1 and a 1.5% RSD at 15 mg Cu L À1 .…”

“…Further advantages of the use of the analytical sensitivity estimated by means of the capability of discrimination are: 21 it employs the same conceptual methodology as ISO and IUPAC for dening the capability of detection; it is possible to compare analytical sensitivities from different methods (even despite they use signals with different units) because they are expressed in concentration units; and, nally, the a and b errors are considered. A major difference of the multivariate analytical sensitivity over classical calibration and NAS is that it does not depend on 'pseudosignals' or on vectorial decomposition.…”

New Ways for the Advanced Quality Control of Liquefied Natural Gas

“…Bias can be, simply, ∑(x pred −x true ) m , (m = number of samples). Repeatability precision can be evaluated quite easily as the standard error of the regression line (s y/x ), just assuming that the error in the reference values is negligible in comparison to the PLS predictions [65,71]. That is true when using standardized chromatographic methodologies but not so evident when complex methods, such as the motor methane number, are used.…”