| Texto completo | |
| Autor(es): |
da Silva, Leandro P.
;
Oliveira, Micael D. L.
;
Villa, Javier E. L.
Número total de Autores: 3
|
| Tipo de documento: | Artigo Científico |
| Fonte: | Food Chemistry; v. 488, p. 10-pg., 2025-10-01. |
| Resumo | |
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 (O-H bonds), protein content (C-H and N-H 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. (AU) | |
| Processo FAPESP: | 24/09412-4 - Sensores espectroquímicos ultrassensíveis aliados à quimiometria: investigação multivariada de perfis metabólicos de fluídos biológicos usando espectroscopia Raman intensificada por superfície |
| Beneficiário: | Javier Erick Lobatón Villa |
| Modalidade de apoio: | Auxílio à Pesquisa - Regular |
| Processo FAPESP: | 21/05679-8 - EMU concedido no processo 2014/50951-4: Instituto Nacional de Ciências e Tecnologias Analíticas Avançadas (INCTAA). Espectrofotômetro de infravermelho próximo (NIR) |
| Beneficiário: | Javier Erick Lobatón Villa |
| Modalidade de apoio: | Auxílio à Pesquisa - Programa Equipamentos Multiusuários |