Application of genetic algorithm and multivariate methods for the detection and measurement of milk‐surfactant adulteration by attenuated total reflection and near‐infrared spectroscopy

Abstract

BACKGROUND The adulteration of milk by hazardous chemicals like surfactants has recently increased. It conceals the quality of the product to gain profit. As milk and milk‐based products are consumed by many people, novel analytical procedures are needed to detect these adulterants. This study focused on Fourier‐transform infrared (FTIR) spectroscopy equipped with an attenuated total reflection (ATR) accessory, and near‐infrared (NIR) spectroscopy for the determination of milk‐surfactant adulteration using a genetic algorithm (GA) coupled with multivariate methods. The model surfactant was sodium dodecyl sulfate (SDS), and its concentration varied from 1.94–19.4 gkg−1 in adulterated samples. RESULTS Prominent peaks in the spectral range of 5500–6400 cm−1, 1160–1260 cm−1 and 1049–1080 cm−1 may correspond to the sulfonate group in SDS. A genetic algorithm could significantly reduce the number of variables to almost one third by selecting the specific wavenumber region. Principal component analysis (PCA) for ATR and NIR data indicated separate clusters of samples in terms of the concentration level of SDS (P ≤ 0.05). Partial least squares regression (PLSR) was used to determine the maximum R2 value for ATR and NIR data for calibration, cross‐validation and prediction, which were 0.980, 0.972, 0.980, and 0.970, 0.937, and 0.956 respectively. The results showed apparent differences between unadulterated and adulterated samples using partial least squares‐discriminant analysis (PLS‐DA), which was validated by the permutation test. CONCLUSION The results clearly show the successful application of the proposed methods with multivariate analysis in the selection of variables, classification, clustering, and identification of the adulterant in amounts as low as 1.94 gkg−1 in milk. © 2020 Society of Chemical Industry

Near‐infrared spectroscopy for the inline classification and characterization of fruit juices for a product‐customized flash pasteurization

“…The characterization and classification of the fruit juices were realized with chemometric methods applying Simca 16.1 (MKS Umetrics AB). An informative overview of the basics of the chemometric methods used here is given by Hosseini et al (2021) in their publication, thus following is a brief explanation of the partial least squares regression (PLSR) and partial least squares discriminant analysis (PLS-DA) methods used here. For characterization through quantification of ingredients, the PLSR method was applied.…”

“…Near‐infrared spectroscopic investigations on fruit juices are the matter of numerous publications. Typical applications are the verification of the correct declaration of regional provenance, the verification of authenticity, or the quantification of special ingredients (Hosseini et al., 2021 ; Igual et al., 2010 ; Kelly & Downey, 2005 ; Lanza & Li, 1984 ; Rambla et al., 1997 ; Reid et al., 2005 ; Šnurkovič, 2013 ; Twomey et al., 1995 ; Włodarska et al., 2018 ). In most cases, these investigations are offline applications and place emphasis on a single fruit variety or on particular ingredients.…”

Qualitative identification of mature milk adulteration in bovine colostrum using noise‐reduced dielectric spectra and linear model

“…Principal component analysis (PCA) is an unsupervised cluster method that identifies different classes of samples based on similarities 28 . Meanwhile, a smaller number of underlying variables, named principal components (PCs), can be projected from the original variables by PCA 27 …”

“…22,25 The Kennard-Stone (K S) algorithm, a widely used method for qualitative identification, 26 was applied to divide the spectral data of colostrum (49 samples) and adulterated colostrum (105 samples) into calibration and validation sets according to the ratio of 3:1. To select the optimal pretreatment method with general applicability, linear model-partial least squares discriminant analysis (PLS-DA) 27 and non-linear model-support vector machine (SVM) 24 were used. As typical linear and non-linear models, respectively, PLS-DA and SVM are widely used for determining the optimal pretreatment method and qualitative identification.…”

“…28 Meanwhile, a smaller number of underlying variables, named principal components (PCs), can be projected from the original variables by PCA. 27 Considering the adverse effect of inhomogeneity of colostrum on the identification based on dielectric spectroscopy, PCA was used to investigate the clustering effect at different doping levels. On the other hand, PCA was applied to reduce the data dimension during model establishment.…”