Combining Multiple Classifiers based on Third-Order Dependency

Authors:
Hee-Joong Kang;David Doermann
Affiliations:
-;-
Venue:
ICDAR '03 Proceedings of the Seventh International Conference on Document Analysis and Recognition - Volume 1
Year:
2003

Citing 2
Cited 0

A Method of Combining Multiple Experts for the Recognition of Unconstrained Handwritten Numerals

IEEE Transactions on Pattern Analysis and Machine Intelligence
Optimal approximation of discrete probability distribution with kth-order dependency and its application to combining multiple classifiers

Pattern Recognition Letters

Quantified Score

Hi-index	0.00

Visualization

Abstract

Without an independence assumption, combining multipleclassifiers deals with a high order probability distributioncomposed of classifiers and a class label. Storing andestimating the high order probability distribution is exponentiallycomplex and unmanageable in theoretical analysis,so we rely on an approximation scheme using the dependency.In this paper, as an extension of the second-order dependencyapproach, the probability distribution is optimallyapproximated by the third-order dependency and multipleclassifiers are combined. The proposed method is evaluatedon the recognition of unconstrained handwritten numeralsfrom Concordia University and the University of California,Irvine. Experimental results support the proposed methodas a promising approach.