Exploiting temporal information in higher-order fluorescence data for enhanced pattern recognition: Application to saffron geographical authentication
Analytica Chimica Acta · 2026-12 · 1427 · 346422
DOI: 10.1016/j.aca.2026.346422 ↗Kurz gefasst
Fachartikel in der Zeitschrift Analytica Chimica Acta, 2026-12. Verfasst von Elena Gago Prieto und 4 weiteren Autor:innen. Beteiligte Einrichtung: Department of Analytical Chemistry, University of Extremadura, Badajoz, 06006, Spain. Hauptthemen laut MEDLINE: Crocus, Pattern Recognition, Automated. Die Arbeit nennt 30 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0). Gefördert von: Extremadura Government, Spain Ministry of Science Innovation and Universities.
Abstract (Originalsprache)
BACKGROUND: Recent studies using fluorescence data have demonstrated that incorporating additional experimental modes can provide valuable discriminatory information. However, the discriminant value of higher-order data may depend not only on the information acquired, but also on how it is extracted and represented before classification. Here, an in-flow photodegradation strategy was developed to generate time-resolved EEM fluorescence data from saffron extracts. Different data structures were evaluated for saffron geographical authentication. Following PARAFAC decomposition, the resulting latent variables were used as inputs for supervised and one-class classification approaches to assess whether preserving sample-specific temporal fluorescence evolution improves pattern recognition performance. RESULTS: Three data structures were evaluated: second-order/three-way, third-order/four-way, and third-order/augmented three-way. PARAFAC decomposition was performed for each structure, and the resulting scores were used as input variables for supervised and one-class classification. In supervised classification using a backpropagation neural network, all three data structures provided good discrimination among saffron geographical classes. However, differences between data representations became more pronounced when DD-SIMCA was applied. All three models achieved 100 % sensitivity for the target class, whereas the percentage of correctly rejected non-target samples increased from 68 % for the second-order/three-way data to 86 % for the third-order/four-way data and 100 % for the augmented three-way data. These results demonstrate that the improvement from higher-order data is not solely due to incorporating an additional temporal dimension, but also to how temporal information is retained in the PARAFAC score matrix and subsequently transferred to the classification model. SIGNIFICANCE: This study demonstrates that data representation is a critical factor when exploiting higher-order fluorescence information. Although all data structures provided good supervised discrimination, preserving sample-specific temporal evolution substantially improved one-class classification. The results highlight the importance of designing dimensionality-reduction strategies that retain system-specific information relevant to class discrimination.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Analytical Chemistry, University of Extremadura, Badajoz, 06006, Spain.
- Laboratorio de Desarrollo Analítica y Quimiometría (LADAQ), Cátedra de Química Analítica I, Facultad de Bioquímica y Ciencias Biológicas, Universidad Nacional del Litoral, Santa Fe, S3000ZAA, Argentina; Consejo Nacional…
- Department of Analytical Chemistry, University of Extremadura, Badajoz, 06006, Spain; Research Institute on Water, Climate Change and Sustainability (IACYS), University of Extremadura, Badajoz, 06006, Spain.
- Department of Analytical Chemistry, University of Extremadura, Badajoz, 06006, Spain; Research Institute on Water, Climate Change and Sustainability (IACYS), University of Extremadura, Badajoz, 06006, Spain. Electronic…
Förderung
- Extremadura Government
- Spain Ministry of Science Innovation and Universities
Zitieren
APA
Gago Prieto, E., Goicoechea, H. C., Muñoz de la Peña, A., Durán Merás, I., & Alcaraz, M. R. (2026). Exploiting temporal information in higher-order fluorescence data for enhanced pattern recognition: Application to saffron geographical authentication. Analytica Chimica Acta, 1427, 346422. https://doi.org/10.1016/j.aca.2026.346422
Vancouver
Gago Prieto E, Goicoechea HC, Muñoz de la Peña A, Durán Merás I, Alcaraz MR. Exploiting temporal information in higher-order fluorescence data for enhanced pattern recognition: Application to saffron geographical authentication. Analytica Chimica Acta. 2026;1427:346422. doi:10.1016/j.aca.2026.346422
BibTeX
@article{GagoPrieto2026,
author = {Gago Prieto, Elena and Goicoechea, Héctor C and Muñoz de la Peña, Arsenio and Durán Merás, Isabel and Alcaraz, Mirta R},
title = {Exploiting temporal information in higher-order fluorescence data for enhanced pattern recognition: Application to saffron geographical authentication},
journal = {Analytica Chimica Acta},
year = {2026},
volume = {1427},
pages = {346422},
doi = {10.1016/j.aca.2026.346422}
}
EUVIMED