Seminars

A general logistic type discriminant analysis with U-divergence for contaminated data

237
reads

Su-Yun Huang

2008-06-10
13:30:00 - 15:00:00

A general logistic type discriminant analysis with U-divergence for contaminated data

405 , Mathematics Research Center Building (ori. New Math. Bldg.)

We will first review works by S. Eguchi et al. on logistic type discriminant analysis and its link to U-divergence, Neyman-Pearson Lemma and area under ROC curve. Next, we will discuss the extension of logistic type discriminant analysis to kernel-based support vector machines for contaminated data.