Title :
An evidence-theoretic neural network classifier
Author :
Denoeux, Thierry
Author_Institution :
Univ. de Technol. de Compiegne
Abstract :
A new classifier based on the Dempster-Shafer theory of evidence is presented. The approach consists in considering the similarity to prototype vectors as evidence supporting certain hypotheses concerning the class membership of a pattern to be classified. The different items of evidence are represented by basic belief assignments over the set of classes and combined by Dempster´s rule of combination. An implementation of this procedure in a neural network with specific architecture and learning procedure is presented. A comparison with LVQ and RBF neural network classifiers is performed
Keywords :
case-based reasoning; learning (artificial intelligence); neural nets; pattern classification; probability; Dempster´s rule of combination; Dempster-Shafer theory; LVQ neural network classifiers; RBF neural network classifiers; basic belief assignments; evidence-theoretic neural network classifier; Computer architecture; Costs; Multi-layer neural network; Neural networks; Pattern classification; Prototypes; Robustness; Training data; Vector quantization; Voting;
Conference_Titel :
Systems, Man and Cybernetics, 1995. Intelligent Systems for the 21st Century., IEEE International Conference on
Conference_Location :
Vancouver, BC
Print_ISBN :
0-7803-2559-1
DOI :
10.1109/ICSMC.1995.537848
