Caracterización y discriminación de parafinas de uso agroalimentario mediante espectroscopía Vis-NIR y aprendizaje automático

  1. Barea-Sepúlveda, M. 1
  2. Ferreiro-González, M. 1
  3. Calle, J.L.P. 1
  4. Palma, M. 1
  1. 1 Departamento de Química Analítica, Facultad de Ciencias, Universidad de Cádiz, Campus Internacional de Excelencia Agroalimentaria (ceiA3); IVAGRO,11510, Puerto Real, Cádiz.
Book:
XLIV JORNADAS DE VITICULTURA Y ENOLOGÍA DE LA TIERRA DE BARROS

Publisher: Centro Universitario Santa Ana

Year of publication: 2022

Pages: 483-502

Type: Book chapter

GOOGLE SCHOLAR

Abstract

Waxes are petroleum-derived products (PDPs) with a wide spectrum of industrial and consumer applications that vary according to their chernical composition. This study presents a method based on visible and near-infrared spectroscopy in combination with machine leaming for the correct characterization and discrimination of the two most marketed types of petroleum waxes. Moreover, the spectroscopic data combined with unsupervised machine leaming algorithms, such as hier archical cluster analysis (HCA), and with nonparametric su pervised machine leaming algorithms, such as support vector machines (SVM) and random forests (RF), allowed to charac terize, and discriminate, the samples based on their molecular composition. The results obta.ined demonstrated the·suitability of this fast, environmentally friendly, and cost-effective analyti cal technique as an altemative to current methods for automatic quality control of petroleum waxes.