Comprehensive and Non-Destructive Sweet Corn Shelf-Life Prediction Using Near-Infrared (NIR) Spectroscopy Coupled with Multivariate Curve Resolution-Alternating Least Squares (MCR-ALS) Spectral Resolution

Accurate shelf-life prediction of perishable products remains challenging because quality deterioration involves multiple physicochemical changes that are not adequately captured by conventional univariate approaches. This study proposes a multivariate shelf-life prediction framework for sweet corn based on near-infrared (NIR) spectroscopy coupled with multivariate curve resolution–alternating least squares (MCR-ALS). NIR spectra were collected from sweet corn samples and analyzed using MCR-ALS to extract chemically interpretable concentration and spectral profiles. A total of 100 and 85 corn samples were used for model training and validation, respectively. The dominant MCR-ALS component showed strong correlations with total soluble solids, dry matter, and individual sugar contents (sucrose, glucose, and fructose), effectively describing the overall quality degradation process. Based on the zero-order kinetic model, the predicted shelf lives were 41.3, 11.0, and 8.9 days at 4, 13, and 25 °C, respectively. Arrhenius analysis of the MCR-ALS concentration profile yielded a temperature-dependent degradation rate with an activation energy of 54.05 kJ mol−1 (R2 = 0.8387). The practical applicability of the proposed framework was further examined using a separate harvest batch of sweet corn that underwent repeated non-destructive NIR measurements throughout storage. Overall, the proposed NIR–MCR-ALS framework provides a rapid, non-destructive, and chemically interpretable approach for shelf-life prediction and postharvest quality monitoring of perishable produce.