Stage-Dependent Fuzzy-valued Loss Function in Two-Stage Binary Classifier
In this paper, a model to deal with two-stage Bayesian classifier, under the assumption of complete probabilistic information, is introduced. The loss function in our problem is stage-dependent fuzzy-valued. This fuzzy loss function means that the loss depends on the stage at which misclassification is made. The model is firstly based on the notion of fuzzy random variable and secondly on the subjective ranking of fuzzy number defined by Campos and González. The comparison with crisp stage-dependent loss function is given. Finally, an example illustrating this case of Bayesian analysis is considered.
KeywordsBayes rule multistage classifier fuzzy loss function
Unable to display preview. Download preview PDF.
- 2.Berger, J.: Statistical Decision Theory and Bayesian Analysis. Springer-Verlag, New York (1993)Google Scholar
- 6.Domingos, P.: MetaCost: A General Method for Making Classifiers Cost-Sensitive. In Proceedings of the Fifth International Conference on Knowledge Discovery and Data Mining (KDD-99), 155–164 (1999)Google Scholar
- 8.Gil, M.A., López-Dìaz, M.: A Model for Bayesian Decision Problems Involving Fuzzy-Valued Consequences. Proc. 6-th Int. Conf. on Information Processing and Management of Uncertainty in Knowledge Based Systems, Granada 495–500 (1996)Google Scholar
- 13.Li-Wei,.K, Bor-Chen K., Chin-Teng L., Der-Jenq, L.: A Two-Stage Classification System with Nearest Neighbor Reject Option for Hyperspectral Image Data, In 18th IPPR Conference on Computer Vision, Graphics and Image Processing (CVGIP) Taipei, 1258–1263 (2005)Google Scholar
- 15.Margineantu, D.D., Dietterich, T.G.: Bootstrap methods for the cost-sensitive evaluation of classifiers, In Proc. 17 Int. Conf. on Machine Learning ICML’2000, Morgan Kaufmann, 583–590, (2000)Google Scholar
- 16.Viertl, R.: Statistical Methods for Non-Precise Data. CRC Press, Boca Raton (1996)Google Scholar