A systematic review of analytical techniques coupled with emerging machine learning and chemometric approaches for detecting yogurt adulteration

Abstract

Yogurt adulteration has escalated globally, posing significant risks to both consumer health and market integrity. This systematic review critically examines the application of key analytical techniques-infrared, Raman, fluorescence, nuclear magnetic resonance, stable isotope analysis, mass spectrometry, and hyperspectral imaging-to detect adulterants such as vegetable oils, non-dairy proteins, and synthetic additives. The distinct contribution of this work lies in its comprehensive integration of advanced chemometric and machine learning algorithms, such as principal component analysis, partial least squares methods, support vector machines, artificial neural networks, and deep learning approaches. These algorithms not only enhance discrimination accuracy but also enable reliable quantification of adulterants at trace concentrations. Following PRISMA guidelines, we included more than thirty-five peer-reviewed studies specifically targeting yogurt adulteration and thoroughly documented their methodologies, analytical performances, and data analysis frameworks. The review highlights that analytical techniques, when combined with chemometric approaches or machine learning models, consistently achieve high sensitivity and specificity. However, challenges remain, including the high cost of instrumentation, the need for extensive calibration, and the complexity of large-scale data interpretation. Emerging trends focus on miniaturizing analytical devices and further integrating deep learning strategies to facilitate real-time and on-site authenticity verification. By elucidating current strengths, limitations, and prospective developments, this review provides a robust foundation for future research aimed at streamlining yogurt adulteration detection.