POSSIBILISTIC METHODOLOGY FOR THE EVALUATION OF CLASSIFICATION ALGORITHMS

Olgierd Hryniewicz

2011

Abstract

In the paper we consider the problem of the evaluation and comparison of different classification algorithms. For this purpose we apply the methodology of statistical tests for the multinomial distribution. We propose to use two-sample tests for the comparison of different classification algorithms, and one-sample goodness-of-fit tests for the evaluation of the quality of classification. We restrict our attention to the case of the supervised classification when an external ‘expert’ evaluates the correctness of classification. The results of the proposed statistical tests are interpreted using possibilistic indices of dominance introduced by Dubois and Prade.

Paper Citation

in Harvard Style

Hryniewicz O. (2011). POSSIBILISTIC METHODOLOGY FOR THE EVALUATION OF CLASSIFICATION ALGORITHMS . In Proceedings of the 6th International Conference on Software and Database Technologies - Volume 2: ICSOFT, ISBN 978-989-8425-77-5, pages 313-322. DOI: 10.5220/0003436803130322

in Bibtex Style

@conference{icsoft11,
author={Olgierd Hryniewicz},
title={POSSIBILISTIC METHODOLOGY FOR THE EVALUATION OF CLASSIFICATION ALGORITHMS},
booktitle={Proceedings of the 6th International Conference on Software and Database Technologies - Volume 2: ICSOFT,},
year={2011},
pages={313-322},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0003436803130322},
isbn={978-989-8425-77-5},
}

in EndNote Style

TY - CONF
JO - Proceedings of the 6th International Conference on Software and Database Technologies - Volume 2: ICSOFT,
TI - POSSIBILISTIC METHODOLOGY FOR THE EVALUATION OF CLASSIFICATION ALGORITHMS
SN - 978-989-8425-77-5
AU - Hryniewicz O.
PY - 2011
SP - 313
EP - 322
DO - 10.5220/0003436803130322