MIDINFRARED SPECTROSCOPY AND CHEMOMETRICS FOR DETERMINING THE TYPE OF FRUIT USED IN JAM

Fourier transform infrared spectra of jams were analysed by chemometric methods with the aim of differentiating between 'strawberry' and 'nonstrawberry' containing jams. Spectra were subjected to a data reduction step, which was either principal component analysis (PCA)-based or partial least squares (PLS)-based, before being classified in one of the two jam types on the basis of the smaller distance to the group 'means'. Diffuse reflectance infrared spectra of the insoluble materials of the jams led to a success in classification of almost 100%. However, the preparation method was lengthy and still needs some improvement. Attenuated total reflectance spectra of jams were classified according to fruit type with 91% success. This method was found to be strongly influenced by the spectral differences between 'normal' and 'reduced' total sugar content jams, which appeared in the first principal component in both the PCA-based and the PLS-based method.