Multicategory Angle-Based Large-Margin Classification

• Activity Number: 203
• Type: Invited
• Date/Time: Monday, August 4, 2014 : 2:00 PM to 3:50 PM
• Sponsor: SSC
• Author(s): Chong Zhang and Yufeng Liu*+
• Companies: University of North Carolina and University of North Carolina at Chapel Hill
• Keywords: Classification ; Large Margin ; Multicategory ; Regularization ; Sum-to-Zero Constraint

Abstract

Large-margin classifiers are popular classification methods in both machine learning and statistics. Despite the success of binary large-margin classifiers, extensions to multicategory problems are quite challenging. Among existing simultaneous multicategory large-margin classifiers, a common approach is to learn k different classification functions for a k-class problem with a sum-to-zero constraint. Such a formulation can be inefficient. In this talk, I will present a new Multicategory Angle-based large-margin Classification (MAC) framework. The proposed MAC structure considers a simplex based prediction rule without the sum-to-zero constraint, and consequently enjoys more efficient computation. Both theoretical and numerical studies will be discussed to demonstrate the usefulness of the proposed MAC classifiers. Extensions for classification of functional data will be discussed as well.