Induction of decision trees from partially classified data using belief functions

Author

Fenoeux, T. ; Bjanger, M. Skarstein

Author_Institution

Univ. de Technol. de Compiegne, France

Volume

4

fYear

2000

fDate

2000

Firstpage

2923

Abstract

A new tree-structured classifier based on the Dempster-Shafer theory of evidence is presented. The entropy measure classically used to assess the impurity of nodes in decision trees is replaced by an evidence-theoretic uncertainty measure taking into account not only the class proportions, but also the number of objects in each node. The resulting algorithm allows the processing of training data whose class membership is only partially specified in the form of a belief function. Experimental results with EEG data are presented

Keywords

belief networks; decision trees; electroencephalography; entropy; learning by example; medical signal processing; pattern classification; uncertainty handling; Dempster-Shafer theory; EEG data; belief functions; class membership; class proportions; decision tree induction; entropy measure; evidence; evidence-theoretic uncertainty measure; node impurity; partially classified data; training data processing; tree-structured classifier; Classification tree analysis; Decision trees; Electroencephalography; Entropy; Impurities; Machine learning; Machine learning algorithms; Measurement uncertainty; Pattern recognition; Training data;

fLanguage

English

Publisher

ieee

Conference_Titel

Systems, Man, and Cybernetics, 2000 IEEE International Conference on

Conference_Location

Nashville, TN

ISSN

1062-922X

Print_ISBN

0-7803-6583-6

Type

conf

DOI

10.1109/ICSMC.2000.884444

Filename

884444