Control charts for chemometric evaluation of Raman spectra

Principal component analysis (PCA) is a method used for evaluation of vibration spectra that allows separating important information from the minor effects. From the point of view of data structure the most valuable information is depicted in score graphs of several Principal Components (PCs), the contribution of various spectral regions can be viewed in the x-loadings charts. Questionable problems of PCA are the identification of outliers for various spectral variables and various PCs and mainly how to compare mutually multiple large spectral data sets. To solve these issues we attempt to introduce control charts in multivariate spectral analysis integrating them with PCA. A utility was developed to provide comparison of new samples (spectral data sets) according to previous set (calibration) for data in individual PCs. Similar spectral data sets are then inside the control limits while the different spectral data exceed the limits. The correlated data in individual PCs can be easily distinguished from dissimilar spectral data.