An efficient workflow for predicting various logging variables using simple machine-learning programs - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2024

An efficient workflow for predicting various logging variables using simple machine-learning programs

Mehdi Serdoun
  • Fonction : Auteur
Elodie Williard
  • Fonction : Auteur
Patrick Ledru
  • Fonction : Auteur
  • PersonId : 1127004
Thomas Obin
  • Fonction : Auteur
Gaétan Milesi
Amber Doney
  • Fonction : Auteur
Anthony Le Beux
  • Fonction : Auteur
  • PersonId : 1036789
J. Mercadier

Résumé

The paper presents simple tools for the prediction of logging variables in uranium exploration using various instrumental data. These tools include i) the prediction of potassic alteration using routine IR-spectroscopy in the 350-2500nm range using Partial Least Squares Regression (PLS-R), ii) the prediction of alteration facies using Visible-NIR (350-1000nm) spectroscopy along with Partial Least Squares – Discriminant Analysis (PLS-DA) and iii) lithostratigraphic units classification using geochemical assays along with a Random Forest Classifier. These tools are associated with an open online repository describing a standard Machine-Learning pipeline for drilling data.
Fichier non déposé

Dates et versions

hal-04623212 , version 1 (25-06-2024)

Identifiants

  • HAL Id : hal-04623212 , version 1

Citer

Mehdi Serdoun, Frédéric Sur, Elodie Williard, Patrick Ledru, Thomas Obin, et al.. An efficient workflow for predicting various logging variables using simple machine-learning programs. Uranium Raw Material for the Nuclear Fuel Cycle: Innovation for Sustaining Future Resources and Production (URAM), International Atomic Energy Agency, May 2023, Vienna, Austria. Paper 118. ⟨hal-04623212⟩
1 Consultations
0 Téléchargements

Partager

Gmail Mastodon Facebook X LinkedIn More