EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Rethinking local spectral modelling: From per-query refitting to model libraries

Leonardo Ramirez-Lopez, Maxime Metz, Matthieu Lesnoff, Claudio Orellano, Estefania Perez-Fernandez, Marcal Plans, Timo Breure, Thorsten Behrens, Raphael Viscarra Rossel, Yi Peng

Analytica Chimica Acta · 2026-09 · 1415 · 345682

DOI: 10.1016/j.aca.2026.345682 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Analytica Chimica Acta, 2026-09. Verfasst von Leonardo Ramirez-Lopez und 9 weiteren Autor:innen. Beteiligte Einrichtung: Imperial College London, Imperial College Business School, South Kensington Campus, London, SW7 2AZ, England, United Kingdom; BUCHI Labortechnik AG, …. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 54 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Eidgenössische Technische Hochschule Zürich, Nuclear Science and Security Consortium, Natural Resources Conservation Service.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

BACKGROUND: This work introduces liblex, a simple yet powerful algorithmic framework to streamline the use of complex spectral libraries. It converts diffuse reflectance spectroscopy (DRS) libraries into libraries of pre-computed localised models (experts) that can be retrieved and combined at prediction time. RESULTS: The liblex algorithmic framework was evaluated in a challenging cross-continental test case. A large and heterogeneous DRS library of North American soils was used to construct a library of predictive models for total carbon (TC), which were then applied to samples from a region in the Democratic Republic of the Congo. In this scenario, liblex achieved high predictive accuracy and produced RMSE values that were competitive with, and in this case lower than, those obtained with the benchmark methods considered, including LOCAL, Cubist, global PLS variants, and a convolutional neural network with transfer learning. This work also shows how liblex intrinsically provides uncertainty proxies via prediction dispersion and supports interpretability through stored regression coefficients and variable-importance profiles. SIGNIFICANCE: The liblex framework addresses several long-standing drawbacks of conventional local modelling methods by replacing per-query model refitting with the retrieval and aggregation of pre-computed localised experts. This design reduces computational burden at prediction time, enables deployment without access to the full reference library, preserves interpretability, and supports robust prediction across heterogeneous domains. The dispersion among retrieved expert predictions also provides an intrinsic, low-cost proxy for sample-specific uncertainty. These features make liblex a scalable, interpretable, and privacy-preserving strategy for operationalising large spectral libraries.

Schlagwörter

Stichwörter der Autor:innen

Cross-domain modellingDiffuse reflectance spectroscopyLocal chemometricsMemory-based learningRetrieval-gated modellingTransfer learning

Autor:innen und Institutionen

  • Imperial College London, Imperial College Business School, South Kensington Campus, London, SW7 2AZ, England, United Kingdom; BUCHI Labortechnik AG, Department of Data Science, Meierseggstrasse 40, Flawil, CH-9230, Swit…
  • Pellenc ST, Applied Research Group, Pertuis, 84120, France; LabCom Aioly, Artificial Intelligence and Optics Laboratory, Montpellier, 34196, France; Institut Méditerranéen de Biodiversité et d'Ecologie marine et contine…
  • SELMET, Univ. Montpellier, CIRAD, INRAE, Institut Agro, Montpellier, 34090, France; Centre de coopération internationale en recherche agronomique pour le développement (CIRAD), UMR SELMET, Montpellier, 34090, France; Ch…
  • BUCHI Labortechnik AG, Department of Data Science, Meierseggstrasse 40, Flawil, CH-9230, Switzerland.
  • European Commission, Joint Research Centre (JRC), Via E. Fermi, 2749, Ispra, 21027, Italy.
  • Swiss Competence Center for Soil, Länggasse 85, Zollikofen, CH-3052, Switzerland.

Förderung

  • Eidgenössische Technische Hochschule Zürich
  • Nuclear Science and Security Consortium
  • Natural Resources Conservation Service

Zitieren

APA

Ramirez-Lopez, L., Metz, M., Lesnoff, M., Orellano, C., Perez-Fernandez, E., Plans, M., Breure, T., Behrens, T., Viscarra Rossel, R., & Peng, Y. (2026). Rethinking local spectral modelling: From per-query refitting to model libraries. Analytica Chimica Acta, 1415, 345682. https://doi.org/10.1016/j.aca.2026.345682

Vancouver

Ramirez-Lopez L, Metz M, Lesnoff M, Orellano C, Perez-Fernandez E, Plans M, et al. Rethinking local spectral modelling: From per-query refitting to model libraries. Analytica Chimica Acta. 2026;1415:345682. doi:10.1016/j.aca.2026.345682

BibTeX
@article{RamirezLopez2026,
  author = {Ramirez-Lopez, Leonardo and Metz, Maxime and Lesnoff, Matthieu and Orellano, Claudio and Perez-Fernandez, Estefania and Plans, Marcal and Breure, Timo and Behrens, Thorsten and Viscarra Rossel, Raphael and Peng, Yi},
  title = {Rethinking local spectral modelling: From per-query refitting to model libraries},
  journal = {Analytica Chimica Acta},
  year = {2026},
  volume = {1415},
  pages = {345682},
  doi = {10.1016/j.aca.2026.345682}
}
Kostenlos starten →