JSM 2015 Preliminary Program

Online Program Home
My Program

Abstract Details

Activity Number: 53
Type: Invited
Date/Time: Sunday, August 9, 2015 : 4:00 PM to 5:50 PM
Sponsor: SSC
Abstract #314378
Title: Angle Breakdown Point for Classification
Author(s): Yufeng Liu*
Companies: The University of North Carolina
Keywords: Breakdown ; Classification ; Kernel ; Large Margin ; Robustness
Abstract:

As an important measurement of robustness, breakdown point has been widely used for regression setting. However this measurement is not applicable for classification problems in many cases. To better describe the robustness of different classification methods, we propose a new criteria, called angle breakdown point. With the help of this new measurement, we study robustness of classification methods for both binary and multicategory classification problems. We consider different combinations of both bounded and unbounded loss functions with linear or kernel learning. These studies provide an useful insight on the behavior of different methods when outliers are presented. Numerical results confirm our theoretical findings.

This talk is based on the joint work with Junlong Zhao at Beihang University and Guan Yu at University of North Carolina.


Authors who are presenting talks have a * after their name.

Back to the full JSM 2015 program





For program information, contact the JSM Registration Department or phone (888) 231-3473.

For Professional Development information, contact the Education Department.

The views expressed here are those of the individual authors and not necessarily those of the JSM sponsors, their officers, or their staff.

2015 JSM Online Program Home