The discrimination analysis of adulterated milk has been carried out based on two-dimensional (2D) infrared correlation spectroscopy along with multivariate methods like kernel orthogonal projection to latent structure (K-OPLS), multi-way partial least squares discriminant analysis (NPLS-DA), and unfolded partial least squares discriminant analysis (PLS-DA). 2D correlation spectroscopy, due to high spectral resolution and good spectral interpretation capabilities, is suitable for the analysis of complex biological data. 64 pure milk samples and 64 adulterated milk samples were measured in the mid-infrared range of 900–1700 cm⁻¹. Then, the synchronous 2D correlation spectra of all samples were calculated in the region of between 900–1200 cm⁻¹ and 1200–1700 cm⁻¹. Finally, the K-OPLS, NPLS-DA, and unfolded PLS-DA models were developed based on the synchronous 2D correlation spectra of adulterated milk and pure milk. The classification accuracy rates of unknown samples for K-OPLS, NPLS-DA, and unfolded PLS-DA models were 95%, 92.5%, and 92.5%, respectively. The results indicated that 2D correlation infrared spectroscopy combined with multivariate methods were feasible and efficient for discrimination of adulterated milk.

An innovative method for the determination of concentration of methanol in white spirit was proposed using two-dimensional near-infrared (NIR) correlation spectroscopy in combination with multivariate calibration. A total of 38 white spirit samples were adulterated with level of methanol varying from 0.1 to 10%. A one-dimensional NIR spectra of all samples were collected. Spectral characteristics of white spirit adulterated with methanol were studied based on one-dimensional NIR spectra and two-dimensional NIR correlation spectra. The origin of 4396 cm⁻¹ peak of methanol was verified in terms of the existence or absence of cross-peaks in synchronous and asynchronous two-dimensional NIR correlation spectra. For comparison, the quantitative analysis models were constructed to determine content of methanol in white spirit using synchronous two-dimensional NIR correlation spectra and traditional one-dimensional NIR spectra, respectively. The prediction statistics criteria for synchronous two-dimensional NIR correlation spectra using N-way partial least squares, average relative error and root mean square error, were 2.97 and 0.064%, respectively. For traditional one-dimensional NIR spectra using partial least squares method, average relative error and root mean square error of prediction were 5.3 and 0.079%, respectively. Comparison results show that the proposed new method has stronger predictive ability and can provide better results than traditional one-dimensional NIR spectra using partial least squares.

A new approach for discrimination of adulterated milk is reported using two-dimensional infrared (IR) correlation spectroscopy by multiway principal component analysis (MPCA) and least squares support vector machines (LS–SVM). First, the synchronous two-dimensional spectra of pure and adulterated milk were calculated. Then, MPCA was used to reduce the dimensions, extract features of two-dimensional correlation data set, and distinguish adulterated milk and pure milk. Finally, a LS-SVM model was developed using the scores of the first thirteen principal components from synchronous two-dimensional correlation spectra computed by MPCA as the input variables. The ratios of correct classification were 100% and 96.3% for calibration set and prediction set, respectively. The area under the receiver operating characteristic curves (ROC) of 0.991 for prediction set was obtained by LS–SVM. The results indicate that two-dimensional correlation infrared spectra combined with MPCA–LS–SVM may be a rapid screening technique for discrimination of adulterated milk with good accuracy.