A comparison of decision tree-based algorithms for food discrimination using vibrational spectroscopy

• Published in: Food chemistry Vol. 488; p. 144909
• Main Authors: da Silva, Leandro P., Oliveira, Micael D.L., Villa, Javier E.L.
• Format: Journal Article
• Language: English
• Published: England Elsevier Ltd 01.10.2025
• Subjects: Algorithms; Bread - analysis; Cocos - chemistry; Decision Trees; Discriminant Analysis; Explainable algorithms; Food Analysis - methods; Food Contamination - analysis; Glutens - analysis; Glutens - chemistry; Infrared spectroscopy; Machine learning; Python; Raman scattering; Spectroscopy, Near-Infrared - methods; Spectrum Analysis, Raman - methods; Infrared spectroscopy; Explainable algorithms; Raman scattering; Machine learning; Python
• ISSN: 0308-8146; 1873-7072; 1873-7072
• DOI: 10.1016/j.foodchem.2025.144909

In this work, we systematically studied decision tree-based algorithms—decision tree, random forest, and XGBoost—for binary food discrimination. Using near-infrared (NIR) and Raman spectroscopy, accuracy values of up to 99 % were achieved for discriminating (1) gluten-containing and gluten-free bread and (2) pure and sucrose-adulterated coconut water, respectively. Moreover, NIR bands of water (OH bonds), protein content (CH and NH bonds), and Raman bands attributed to C–O–C bonds in the glycosidic structure were identified as the most important. In addition to traditional feature importance estimates, a strategy based on impurity reduction was proposed to improve chemical interpretability. The figure of merit and splitting method selected for optimizing the algorithms were also evaluated. Although random forest required a higher computational cost than decision tree and partial least squares discriminant analysis, it outperformed both in accuracy. It also provided more robust and chemically meaningful results than XGBoost. [Display omitted] •Direct bread and coconut water analyses by vibrational spectroscopy.•XGBoost and random forest outperform partial least squares discriminant analysis.•Consistent variable importance estimation approach by using random forest.•Improved chemical interpretability for binary food discrimination.•Systematic model validation by estimating figures of merit.