<![CDATA[Milk adulteration has become a significant issue in Indonesia, particularly in high-value commodities such as Etawa goat milk and its fermented products, which are widely produced in the Special Region of Yogyakarta. Formalin adulteration in fermented dairy products poses a critical food safety concern, necessitating rapid, reliable, and non-destructive analytical strategies. This study aimed to develop a combined fingerprinting FTIR profile and chemometric approach for detecting and quantifying formalin adulteration in kefir across concentrations ranging from 1% to 10% v/v. Spectral data were processed using multivariate modeling, including Linear Discriminant Analysis (LDA) for classification and Principal Component Regression (PCR) and Partial Least Squares (PLS) for quantitative prediction. Linear Dicriminant Analysis (LDA) applied to classify samples and identify unique spectral markers. The results demonstrated that authentic Etawa goat milk exhibited a consistent spectral pattern, while adulterated samples showed significant variations, particularly in protein, methanol and carbohydrate-associated bands. The LDA model successfully distinguished pure and adulterated samples, demonstrating clear class separation driven by formalin-induced spectral variations. Quantitative analysis revealed that the PCR model outperformed PLS, exhibiting superior precision with RMSEC and RMSEP values below 2% and strong linearity (R² > 0.99) within the selected wavenumber region (3600–1100 cm⁻¹). These findings indicate that PCR offers robust predictive capability for formalin detection and can serve as an effective tool for routine quality control. Overall, the integration of FTIR spectroscopy and chemometric modeling provides a rapid, accurate, and reproducible method for monitoring adulteration in fermented dairy products.]]>
Copyrights © 2026
