N. Majcen et al., LINEAR AND NONLINEAR MULTIVARIATE-ANALYSIS IN THE QUALITY-CONTROL OF INDUSTRIAL TITANIUM-DIOXIDE WHITE PIGMENT, Analytica chimica acta, 348(1-3), 1997, pp. 87-100
In order to establish an adequate analytical system for the quality co
ntrol of industrially produced titanium dioxide white pigments, two mu
ltivariate linear calibration techniques, principal component regressi
on (PCR) and partial least squares (PLS), are used to model the relati
onship between the important pigment property, change of colour, and i
ts chemical composition. The results, in terms of accuracy, precision,
suitability for quality control and analysis time are compared to tho
se obtained with artificial neural networks (ANNs). Two multivariate d
isplay techniques, principal component analysis (PCA) and corresponden
ce factor analysis (CFA) together with two hierarchical clustering tec
hniques, divisive and Ward's agglomerative hierarchical clustering, ar
e also applied to the X-ray fluorescence data of the pigments samples
so as to extract as much information as possible. Correlation coeffici
ents obtained by PCR and PLS are 0.92 and 0.94, respectively. Both of
them are higher than the already achieved correlation coefficient by A
NNs [I], but the precision of the model derived by ANNs is better. It
should also be pointed out that some important additional information
about the relations between independent variables (chemical compositio
n of the pigment samples) and about the influence of different oxide c
oncentrations on the pigment property, which could be used in the cont
rolling of the production process, was found out.