Calibrating Margin-Based Classifier Scores into Polychotomous Probabilities
作者： Martin Gebel , Claus Weihs
DOI: 10.1007/978-3-540-78246-9_4
关键词:
摘要: Margin-based classifiers like the SVM and ANN have two drawbacks. They are only directly applicable for two-class problems they output scores which do not reflect assessment uncertainty. K-class probabilities usually generated by using a reduction to binary tasks, univariate calibration further application of pairwise coupling algorithm. This paper presents an alternative with usage Dirichlet distribution.
springer.com
sci-hub.st 参考文章(7)
Martin Gebel, Claus Weihs, Calibrating Classifier Scores into Probabilities GfKl. pp. 141- 148 ,(2007) , 10.1007/978-3-540-70981-7_17
J.A.K. Suykens, J. Vandewalle, Least Squares Support Vector Machine Classifiers Neural Processing Letters. ,vol. 9, pp. 293- 300 ,(1999) , 10.1023/A:1018628609742
Morris H. DeGroot, Stephen E. Fienberg, The Comparison and Evaluation of Forecasters. The Statistician. ,vol. 32, pp. 12- 22 ,(1983) , 10.2307/2987588
Trevor Hastie, Robert Tibshirani, Classification by pairwise coupling Annals of Statistics. ,vol. 26, pp. 451- 471 ,(1998) , 10.1214/AOS/1028144844
Tong Zhang, Statistical behavior and consistency of classification methods based on convex risk minimization Annals of Statistics. ,vol. 32, pp. 56- 134 ,(2003) , 10.1214/AOS/1079120130
Reducing multiclass to binary: a unifying approach for margin classifiers Journal of Machine Learning Research. ,vol. 1, pp. 113- 141 ,(2001) , 10.1162/15324430152733133
Reinhold Decker, Hans J. Lenz, Advances In Data Analysis ,(2007)