SVM-based classifier design with controlled confidence
Meiting Li, Ishwar K. Sethi · Proceedings of the 17th International Conference on Pattern Recognition, 2004. ICPR 2004. · 2004
A new classification methodology with controlled error rates and a reject option is proposed in this paper. The proposed methodology is implemented using support vector machine's (SVM's) posterior probability preserving property. A new nonparametric method is proposed to accurately estimate error rates from the output of a trained SVM. The experimental results clearly demonstrate the efficacy of the suggested classifier design methodology.