New approaches in the chemometric analysis of infrared spectra of extra-virgin olive oils

The aim of this paper is to apply new chemometric approaches to obtain quantitative information from near and mid infrared spectra of Andalusian extra-virgin olive oils, using gas chromatography as a classical reference analytical technique. Estimations of the content in saturated, monounsaturated a...

Descripción completa

Detalles Bibliográficos
Autores: Sánchez-Rodríguez, María Isabel, Sánchez-López, Elena M., Marinas, Alberto, Caridad, José Mª, Urbano, Francisco José, Marinas, José Mª
Tipo de recurso: artículo
Fecha de publicación:2014
País:España
Institución:Universitat Politècnica de Catalunya (UPC)
Repositorio:UPCommons. Portal del coneixement obert de la UPC
Idioma:inglés
OAI Identifier:oai:upcommons.upc.edu:2117/88564
Acceso en línea:https://hdl.handle.net/2117/88564
Access Level:acceso abierto
Palabra clave:Extra-virgin olive oil
infrared spectroscopy
partial least squares regression
cross-validation
Classificació AMS::62 Statistics::62H Multivariate analysis
Classificació AMS::62 Statistics::62J Linear inference, regression
Classificació AMS::62 Statistics::62Q05 Statistical tables
Àrees temàtiques de la UPC::Matemàtiques i estadística::Estadística matemàtica
Descripción
Sumario:The aim of this paper is to apply new chemometric approaches to obtain quantitative information from near and mid infrared spectra of Andalusian extra-virgin olive oils, using gas chromatography as a classical reference analytical technique. Estimations of the content in saturated, monounsaturated and polyunsaturated fatty acids are given using partial least squares regression from the near and mid infrared data matrices as well as their concatenated matrix. The different estimations are evaluated in terms of goodness of fit (calibration) and prediction (validation), as a function of the number of partial least squares factors in the regression model and the used matrix of data. Furthermore, the nature, systematic or random, of the prediction errors is studied by a decomposition of their mean squared error. Finally, procedures of cross-validation are implemented in order to generalize the previous results.