Fecha de publicación: -- Fuente: PubMed "olive oil" Food Chem. 2026 Oct 13;532(Pt 2):151546. doi: 10.1016/j.foodchem.2026.151546. Online ahead of print.ABSTRACTTransitioning toward high-value functional food markets increases demand for forensic tools capable of verifying the "water-nutrient nexus", the complex agronomic interplay between irrigation and fertilization that shapes olive oil composition and authenticity. Conventional chemometrics often fail to capture synergistic, non-linear cultivation fingerprints within complex chemical matrices. This study introduces a "Deep Chemometrics" framework for retrospective auditing of olive oil history using FTIR spectroscopy and a customized one-dimensional convolutional neural network (1D-CNN). The framework was developed from 324 FTIR spectra representing nine irrigation × fertilization (W × F) regimes and incorporated physics-informed data augmentation to overcome limited sample availability. The optimized 1D-CNN achieved 97.89% validation accuracy (validation loss = 0.0548) and 83% accuracy on a held-out test set (macro-averaged precision = 0.85, recall = 0.83, F1-score = 0.83), representing a 25.8% improvement over the baseline model. Gradient-based Explainable Artificial Intelligence (XAI) consistently identified the ester carbonyl stretching vibration (1744 cm-1) and CH₂ asymmetric stretching vibration (2922 cm-1) as the principal spectral features driving classification. Representative constrained pseudo-Voigt peak deconvolution, together with non-parametric (Kruskal-Wallis) statistical analysis, confirmed significant treatment-dependent differences within these diagnostic bands (1744 cm-1: p = 0.0076; 2922 cm-1: p = 0.0025), providing independent spectroscopic validation of the CNN-derived feature importance. By integrating deep learning, explainable artificial intelligence, and spectroscopic validation, the proposed framework provides a transparent, scientifically interpretable, and robust strategy for retrospective authentication of olive oils according to their irrigation and fertilization history, thereby strengthening agricultural transparency and supporting supply-chain authenticity.PMID:42887244 | DOI:10.1016/j.foodchem.2026.151546