An R package for hybrid uncertainty analysis in natural and environmental risk assessments using probability, imprecise probability and possibility distributions - HAL-SHS - Sciences de l'Homme et de la Société Accéder directement au contenu
Communication Dans Un Congrès Année : 2018

An R package for hybrid uncertainty analysis in natural and environmental risk assessments using probability, imprecise probability and possibility distributions

Résumé

Uncertainty analysis is an unavoidable risk assessment task (for instance for natural hazards, or for environmental issues). In situations where data are scarce, incomplete or imprecise, the systematic and only use of probabilities can be debatable. Over the last years, several alternative mathematical representation methods have been developedto handle in a more flexible manner the lack of knowledge related to input parameters of risk assessment models. This article presents an R package HYRISK dedicated to jointly handling different mathematical representation tools, namely probabilities, possibility distributions and probability functions with imprecise parameters, for the different stages of uncertainty treatment in risk assessments (i.e. uncertainty representation, propagation, sensitivity analysis and decision-making). We support the description using the case study of a dike stability analysis. The package is available at: https://cran.r-project.org/web/packages/HYRISK/index.html.
EGU2018-6208.pdf (38.76 Ko) Télécharger le fichier

Dates et versions

hal-01863418 , version 1 (28-08-2018)

Identifiants

  • HAL Id : hal-01863418 , version 1

Citer

Jeremy Rohmer, Jean-Charles Manceau, Dominique Guyonnet, Faïza Boulahya, Didier Dubois. An R package for hybrid uncertainty analysis in natural and environmental risk assessments using probability, imprecise probability and possibility distributions. 20th European Geosciences Union General Assembly (EGU 2018), Apr 2018, Vienne, Austria. pp.6208. ⟨hal-01863418⟩
128 Consultations
10 Téléchargements

Partager

Gmail Facebook X LinkedIn More