Loading…

Geographical identification of saffron (Crocus sativus L.) by linear discriminant analysis applied to the UV–visible spectra of aqueous extracts

[Display omitted] •We attempted geographical classification of Italian saffron using UV–visible spectra.•Useful regions of the spectrum were found by step-wise linear discriminant analysis.•The quality category of the spices did not influence classification performance.•Correct predictions were 83%...

Full description

Saved in:
Bibliographic Details
Published in:Food chemistry 2017-03, Vol.219, p.408-413
Main Authors: D’Archivio, Angelo Antonio, Maggi, Maria Anna
Format: Article
Language:English
Subjects:
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:[Display omitted] •We attempted geographical classification of Italian saffron using UV–visible spectra.•Useful regions of the spectrum were found by step-wise linear discriminant analysis.•The quality category of the spices did not influence classification performance.•Correct predictions were 83% for 81 samples with 50% of data in the prediction set. We attempted geographical classification of saffron using UV–visible spectroscopy, conventionally adopted for quality grading according to the ISO Normative 3632. We investigated 81 saffron samples produced in L’Aquila, Città della Pieve, Cascia, and Sardinia (Italy) and commercial products purchased in various supermarkets. Exploratory principal component analysis applied to the UV–vis spectra of saffron aqueous extracts revealed a clear differentiation of the samples belonging to different quality categories, but a poor separation according to the geographical origin of the spices. On the other hand, linear discriminant analysis based on 8 selected absorbance values, concentrated near 279, 305 and 328nm, allowed a good distinction of the spices coming from different sites. Under severe validation conditions (30% and 50% of saffron samples in the evaluation set), correct predictions were 85 and 83%, respectively.
ISSN:0308-8146
1873-7072
DOI:10.1016/j.foodchem.2016.09.169