Contextual and isolated algorithms for multistage pattern recognition

Authors:
Marek Kurzynski
Affiliations:
Chair of Systems and Computer Networks, Faculty of Electronics, Wroclaw University of Technology, Wroclaw, Poland
Venue:
SMO'07 Proceedings of the 7th WSEAS International Conference on Simulation, Modelling and Optimization
Year:
2007

Citing 5
Cited 0

On the multistage Bayes classifier

Pattern Recognition
Leave-One-Out Procedures for Nonparametric Error Estimates

IEEE Transactions on Pattern Analysis and Machine Intelligence
Dynamic Programming and Stochastic Control

Dynamic Programming and Stochastic Control
Pattern Classification (2nd Edition)

Pattern Classification (2nd Edition)
Patterns in pattern recognition: 1968-1974

IEEE Transactions on Information Theory

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper deals with the recognition algorithms of a multistage classifier based on a decision tree scheme. For the given tree skeleton and features to be used, concepts of contextual and isolated decision rules (strategies) for performing the classification are discussed. These both strategies are compared in respect of classification accuracy and the upper bound of difference between their probabilities of misclassification is given. The empirical versions of contextual and isolated strategies were practically implemented in the computer-aided prognosis of sacroileitis development and results of classification accuracy on the real data are presented.