Date(s) - 19/03/2018
15 h 30 min - 16 h 30 min
Catégories Pas de Catégories
In machine learning, support vector machines (SVMs) are supervised learning models with associated learning algorithms that analyze data used for classification and regression analysis. In this talk, I will introduce a new margin multi category classifier based on classes of vector valued functions with one component function per category, it is a kernel machine whose separation surfaces are hyperbolic and generalizes the SVMs. I will also exhibit the statistical properties of this classifier , I will show that the classes of component functions are uniform Glivenko-Cantelli (GC). I will then found the guaranteed risk of this classifier.
Finally, I will exhibit a margin loss function ensuring the Fisher consistency.