Improvement of the Identification of Four Heavy Metals in Environmental Samples by Using Predictive Decision Tree Models Coupled with a Set of Five Bioluminescent Bacteria - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue Environmental Science and Technology Année : 2011

Improvement of the Identification of Four Heavy Metals in Environmental Samples by Using Predictive Decision Tree Models Coupled with a Set of Five Bioluminescent Bacteria

Sulivan Jouanneau
Marie-José Durand
Thomas Blusseau
  • Fonction : Auteur
Gerald Thouand

Résumé

A primary statistical model based on the crossings between the different detection ranges of a set of five bioluminescent bacterial strains was developed to identify and quantify four metals which were at several concentrations in different mixtures: cadmium, arsenic III, mercury, and copper. Four specific decision trees based on the CHAID algorithm (CHi-squared Automatic Interaction Detector type) which compose this model were designed from a database of 576 experiments (192 different mixture conditions). A specific software, 'Metalsoft', helped us choose the best decision tree and a user-friendly way to identify the metal. To validate this innovative approach, 18 environmental samples containing a mixture of these metals were submitted to a bioassay and to standardized chemical methods. The results show on average a high correlation of 98.6% for the qualitative metal identification and 94.2% for the quantification. The results are particularly encouraging, and our model is able to provide semiquantitative information after only 60 min without pretreatments of samples.
Fichier non déposé

Dates et versions

hal-01929275 , version 1 (21-11-2018)

Identifiants

Citer

Sulivan Jouanneau, Marie-José Durand, Philippe Courcoux, Thomas Blusseau, Gerald Thouand. Improvement of the Identification of Four Heavy Metals in Environmental Samples by Using Predictive Decision Tree Models Coupled with a Set of Five Bioluminescent Bacteria. Environmental Science and Technology, 2011, 45 (7), pp.2925 - 2931. ⟨10.1021/es1031757⟩. ⟨hal-01929275⟩
73 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More