Apprentissage d’intégrales de Sugeno à partir de données inconsistantes

Quentin Brabant 1 Miguel Couceiro 1
1 ORPAILLEUR - Knowledge representation, reasonning
Inria Nancy - Grand Est, LORIA - NLPKD - Department of Natural Language Processing & Knowledge Discovery
Abstract : The basic setting of this article is multicriteria decision making and preference aggregation. The problem treated is that of learning a Sugeno integral from inconsistent data, where values are elements of a totally ordered set. This is a difficult optimization problem : indeed, a Sugeno integral is determined by 2^n values, with n being the number pf parameters. In this article we propose two learning methods : the first one is an application of simulated annealing, and the second is a new algorithm which relies on the selection of a consistant subset of data and for which the value of n doesn't affect the running time significantly.
Complete list of metadatas

Cited literature [10 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-01404567
Contributor : Quentin Brabant <>
Submitted on : Monday, November 28, 2016 - 8:55:48 PM
Last modification on : Tuesday, December 18, 2018 - 4:38:02 PM
Long-term archiving on : Tuesday, March 21, 2017 - 1:07:54 PM

File

Brabant_Couceiro_LFA2016.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01404567, version 1

Citation

Quentin Brabant, Miguel Couceiro. Apprentissage d’intégrales de Sugeno à partir de données inconsistantes. 25èmes Rencontre Francophone sur la Logique Floue et ses Applications, Nov 2016, La Rochelle, France. pp.49-56. ⟨hal-01404567⟩

Share

Metrics

Record views

271

Files downloads

174