A novel approach based on multiple criteria decision aiding methods to cope with classification problems / A Novel Approach Based on Multiple Criteria Decision Aiding Methods to Cope with Classification Problems

AUTOR(ES)
DATA DE PUBLICAÇÃO

2009

RESUMO

This work presents a novel decision making approach developed upon two Multicriteria Decision Aiding (MCDA) classification methods. The classification methods are centered on the concept of prototypes, that are alternatives that serve as class representatives related to a given problem, and have their performance index very dependent upon the choice of values of some control parameters. In contrast with most MCDA classification research, which put more emphasis on the development of new methods, this study investigates the impact that the prototype selection task exerts on the classification performance exhibited by the classifiers under analysis. To overcome this problem and leverage the classifier performance, two techniques, one based on ELECTRE IV methodology and the other on a customized genetic algorithm, are employed in order to select the prototypes and calibrate the control parameters. To validate the devised classification approach, a case study based on the Alzheimer s disease (AD) was developed to assist clinicians and researchers in the AD early detection. Moreover, the experiments realized, involving different datasets, reveal that there are still some gaps in existing MCDA classification methods that can lead to an improvement of the methodology for the classification problems. Keywords: Multicriteria Decision Aiding, Classification, ELECTRE IV , Genetic Algorithm, Alzheimer s Disease.

ASSUNTO(S)

doenÇa de alzheimer - dissertaÇÕes tomada de decisÃo - dissertaÇÕes anÁlise multicritÉrio - dissertaÇÕes sistemas de informacao algorÍtmos genÉticos - dissertaÇÕes

ACESSO AO ARTIGO

Documentos Relacionados