Séminaire (Organisé par l’Equipe de recherche DI)

Lianmeng JIAO

Heudiasyc/Northwestern Polytechnical University, China

Decision making and pattern classification using belief functions theory

Mardi 17 décembre 2013 à 14h en salle A108

Résumé :

This talk mainly focuses on the applications of belief functions theory in multi-attribute decision making (MADM) and pattern classification. Great uncertainties may exist in MADM problems, such as quantitative and qualitative attributes, missing information in attribute assessment, et al. Belief functions theory, as a general extension of Bayesian theory, is a common and efficient method to solve the high-level fusion problems with uncertainty. We developed a general model of combining sources of evidence with reliability and importance for decision making in belief function framework.

In pattern classification, fuzzy rule-based classification system (FRBCS) is a useful tool. While FRBCS cannot address the imprecise or incomplete information effectively in the modeling and reasoning processes. We extended the FRBCS with belief functions theory for classification applications.


FR SHIC 3272

Collegium UTC/CNRS