Brown rice authenticity evaluation by spark discharge-laser-induced breakdown spectroscopy

PÉREZ-RODRÍGUEZ, MICHAEL; DIRCHWOLF, PAMELA MAIA; SILVA, TIAGO VARÃO; VILLAFAÑE, ROXANA NOELIA; NETO, JOSÉ ANCHIETA GOMES; PELLERANO, ROBERTO GERARDO; FERREIRA, EDILENE CRISTINA

FOOD CHEMISTRY

ELSEVIER SCI LTD

Lugar: Amsterdam; Año: 2019 vol. 297 p. 124960 - 124960

0308-8146

Rice is themost consumed food worldwide, therefore its designation of origin (PDO) is veryuseful. Laser-induced breakdown spectroscopy (LIBS) is an interestinganalytical technique for PDO certification, since it provides fastmultielemental analysis requiring minimal sample treatment. In this work LIBSspectral data from rice analysis were evaluated for PDO certification ofArgentine brown rice. Samples from two PDOs were analyzed by LIBS coupled tospark discharge. The selection of spectral data was accomplished by extremegradient boosting (XGBoost), an algorithm currently used in machine learning,but rarely applied in chemical issues. Emission lines of C, Ca, Fe, Mg and Nawere selected, and the best performance of classification were obtained usingk-nearest neighbor (k-NN) algorithm. The developed method provided 84% ofaccuracy, 100% of sensitivity and 78% of specificity in classification of testsamples. Furthermore, it is simple, clean and can be easily applied for ricecertification.