This is the program for the 2010 Joint Statistical Meetings in Vancouver, British Columbia.

Abstract Details

Activity Number: 112
Type: Topic Contributed
Date/Time: Monday, August 2, 2010 : 8:30 AM to 10:20 AM
Sponsor: Section on Statistical Learning and Data Mining
Abstract - #307763
Title: Large-Margin, Single-Path Hierarchical Classification with a General Loss
Author(s): Lingzhou Xue*+ and Xiaotong Shen
Companies: University of Minnesota and University of Minnesota
Address: 313 Ford Hall, Minneapolis, MN, 55455,
Keywords: Directed acyclic graph ; Hierarchical classification ; Generalization Error ; Structured learning ; Difference convex programming ; Functional genomics

Large margin classifiers deliver high performances for flat classification, but deteriorate dramatically in hierarchical classification. To meet the challenge, we propose a unified large margin approach for single-path hierarchical classification. The key difference with existing approaches is that we consider a general loss to capture the partial correctness rather than complete correctness in generalization, which contains more information and usually preferable in hierarchical classification. Utilizing difference convex programming, we implement the proposed method for support vector machine and psi-learning to obtain hierarchical classifiers. As suggested in both theoretical and numerical analyses, they achieve desired objectives, and outperform other competitors. Finally, our proposed hierarchical classifiers are applied to gene function prediction with gene expression data.

The address information is for the authors that have a + after their name.
Authors who are presenting talks have a * after their name.

Back to the full JSM 2010 program

2010 JSM Online Program Home

For information, contact or phone (888) 231-3473.

If you have questions about the Continuing Education program, please contact the Education Department.