A novel chemometric approach for classifying whole fruit juices, reconstituted juices, and nectars

Fruit juices are widely enjoyed beverages, valued for their organoleptic qualities, chemical composition, and nutritional benefits. They are commonly sold as whole juices (WJs), reconstituted juices, or nectars, with WJs particularly favored by consumers seeking healthier, more balanced dietary options, often commanding a higher market value. However, these desirable characteristics make them vulnerable to fraud and adulteration. To address this issue, a novel strategy has been proposed to classify juices into two categories: WJs and nectar/reconstituted juices (NERJs). This approach combines key chemical markers––such as reducing sugars (RS), sodium (Na), and potassium (K)––with advanced chemometric techniques, including unsupervised methods like principal component analysis (PCA) and supervised methods, such as machine learning algorithms (random forest [RF] and extreme gradient boosting (XGBoost)). The models developed demonstrated remarkable accuracy, achieving 100% classification accuracy for WJs and good performance for NERJs, with accuracy rates > 91.7%. This fast, accurate, and easily implementable strategy is vital for the quality control and authentication of fruit juices by regulatory bodies. It provides essential tools for detecting fraud and adulteration in the beverage industry, ultimately helping ensure the quality and consistency of products available to consumers.