COMMISSION ON GENERAL ASPECTS OF ANALYTICAL CHEMISTRY
Title: Guidelines for Calibration in Analytical Chemistry. Part
2: Multicomponent Calibration.
Danzer, M. Otto and
L. A. Currie
Completion Date: 2004 - Project completed
Part 1 of this project (project 510/25/91) covered Fundamentals
and Single Component calibration [ Pure
70(4), 993-1014, 1998]. The continuation of this first project is
prepared to establish a uniform terminology, notation and formulation
for multicomponent analysis based on multiwavelength spectrometry and/or
multivariate calibration. The main part is dedicated to linear calibration
methods based on unbiased and biased regression methods, such as ordinary
least squares (OLS), principal component regression (PCR) or partial
least squares regression (PLS). Validation of the calibration model,
its predictive performance and the detection of outliners and of influential
observations is carried out by common methods of regression diagnostics.
Some general considerations are made with respect to non-linear multivariate
calibration, such as the method of alternating conditional expectations
(ACE) and neural networks.
The first draft will be prepared before March 1998, sent to the
Commission Secretary to sent to Commission Members for comments to be
discussed at the meeting in August/September 1998.
In December 2002, a final draft was submitted for publication in Pure
and Applied Chemistry.
Project completed - IUPAC Technical Report published in Pure
76(6), 1215-1225, 2004
Last update: 28 July 2004