3.3. Classification of Fibres on the Basis of PCA Analysis of ATR-FT-IR Spectra
Classification of fibres was carried out with Thermo Scientific TQ Analyst Pro software using discriminant analysis which is based on principal component analysis [12].
3.3.1. Classification of Pure Single-component Fibres
Altogether 89 spectra of 11 most common textile fibre classes were used for creating the PCA model and to perform the discriminant analysis. With this model for one component analysis the variances described by the first three principal components were: PC1 45%; PC2 31% and PC3 13%. When illustrating the results with PCA plots, only three principal components are used and it can be seen that some classes significantly overlap. Discriminant analysis can use more than three principal components and this is why with discriminant analysis it is possible to distinguish classes that are superimposed in PCA graph. PCA graph can be seen on Fig. 2.