New data preprocessing trends based on ensemble of multiple preprocessing techniques - IRSTEA - Institut national de recherche en sciences et technologies pour l'environnement et l'agriculture (<b>anciennement Cemagref</b>) Accéder directement au contenu
Article Dans Une Revue Trends in Analytical Chemistry Année : 2020

New data preprocessing trends based on ensemble of multiple preprocessing techniques

Résumé

Data generated by analytical instruments, such as spectrometers, may contain unwanted variation due to measurement mode, sample state and other external physical, chemical and environmental factors. Preprocessing is required so that the property of interest can be predicted correctly. Different correction methods may remove specific types of artefacts while still leaving some effects behind. Using multiple preprocessing in a complementary way can remove the artefacts that would be left behind by using only one technique. This article summarizes the recent developments in new data preprocessing strategies and specifically reviews the emerging ensemble approaches to preprocessing fusion in chemometrics. A demonstration case is also presented. In summary, ensemble preprocessing allows the selection of several techniques and their combinations that, in a complementary way, lead to improved models. Ensemble approaches are not limited to spectral data but can be used in all cases where preprocessing is needed and identification of a single best option is not easily done.
Fichier principal
Vignette du fichier
Mishra2_2020.pdf (1.87 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-02968798 , version 1 (16-10-2020)

Licence

Paternité - Pas d'utilisation commerciale - Pas de modification

Identifiants

Citer

Puneet Mishra, Alessandra Biancolillo, Jean-Michel Roger, Federico Marini, Douglas N Rutledge. New data preprocessing trends based on ensemble of multiple preprocessing techniques. Trends in Analytical Chemistry, 2020, 132, pp.116045. ⟨10.1016/j.trac.2020.116045⟩. ⟨hal-02968798⟩
196 Consultations
375 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More