PLS, iPLS, GA-PLS models for soluble solids content, pH and acidity determination in intact dovyalis fruit using near-infrared spectroscopy
Nenhuma Miniatura disponível
Data
2018-12-01
Autores
de Assis, Mateus WD
De Fusco, Deborah O [UNESP]
Costa, Rosangela C
de Lima, Kássio MG
Cunha Júnior, Luis C
de Almeida Teixeira, Gustavo H [UNESP]
Título da Revista
ISSN da Revista
Título de Volume
Editor
Resumo
BACKGROUND: Dovyalis species Dovyalis abyssinica Warb. and Dovyalis hebecarpa Warb. were introduced into Brazil, but the fruit quality of these species is not appropriate for fresh consumption due to their high titratable acidity (TA) and low soluble solids content (SSC). With the selection of new D. abyssinica clones with lower acidity and the hybridization of these two dovyalis species (D. abyssinica and D. hebecarpa) the fruit quality improved and the better physical–chemical characteristics make them more suitable for fresh consumption. The objective of this study was to develop partial least squares (PLS) models using near infrared spectroscopy (NIRS) for the determination of SSC, TA and pH in intact dovyalis hybrid fruit (D. abyssinica Warb. × D. hebecarpa Warb.). RESULTS: The best SSC prediction model was developed with PLS regression (root mean square error of prediction (RMSEP) of 0.71 °Brix, prediction data set (RP 2) of 0.74 and residual predictive deviation (RPD) of 2.82). Although interval PLS was tested, genetic algorithm PLS performed better for TA (RMSEP of 4.8 g kg−1, RP 2 of 0.40, and RPD of 1.67), and for pH (RMSEP of 0.03, RP 2 of 0.90, and RPD of 6.67). CONCLUSION: NIRS can be used as a non-destructive method to determine quality parameters in intact dovyalis hybrid fruit. © 2018 Society of Chemical Industry.
Descrição
Palavras-chave
chemometrics, D. hebecarpa Warb., Dovyalis abyssinica Warb., NIR, PLS
Como citar
Journal of the Science of Food and Agriculture, v. 98, n. 15, p. 5750-5755, 2018.