Distance Weighted Discrimination

J. S. Marron (marronemail.unc.edu)
M. J. Todd (miketoddcs.cornell.edu)

Abstract: High Dimension Low Sample Size statistical analysis is becoming increasingly important in a wide range of applied contexts. In such situations, it is seen that the popular Support Vector Machine suffers from ``data piling'' at the margin, which can diminish generalizability. This leads naturally to the development of Distance Weighted Discrimination, which is based on Second Order Cone Programming, a modern computationally intensive optimization method.

Keywords: two-class discrimination, second-order cone programming

Category 1: Applications -- Science and Engineering (Statistics )

Category 2: Linear, Cone and Semidefinite Programming (Second-Order Cone Programming )

Citation: Technical Report No. 1339, School of Operations Research and Industrial Engineering, Cornell University (July 2002).

Download: [PDF]

Entry Submitted: 07/25/2002
Entry Accepted: 07/25/2002
Entry Last Modified: 11/29/2004

Modify/Update this entry