DD-Classifiers: New Nonparametric Classification Procedures

• Activity Number: 320
• Type: Invited
• Date/Time: Tuesday, August 3, 2010 : 10:30 AM to 12:20 PM
• Sponsor: IMS
• Author(s): Regina Liu*+ and Jun Li and Juan A. Cuesta-Albertos
• Companies: Rutgers University and University of California, Riverside and University of Cantabria
• Keywords: DD-Classifier ; CLassification ; Multivariate Classification ; nonparametric classification ; DD-plot ; Data Depth

Abstract

Most existing classification algorithms assume either certain parametric distributions for the data or some forms of separating surfaces, which can limit considerably their applicability. We introduce a novel classification algorithm using the so-called DD-plot. This classifier is completely nonparametric, requiring no prior knowledge of the underlying distributions or of the form of the separating surface, and hence has broader applicability. Its classification outcome can be visualized on a two-dimensional plot regardless of the dimension of the data. The DD-classifier is shown to be asymptotically equivalent to the Bayes rule under suitable conditions. The performance of the classifier is examined using simulated and real data sets. Overall, the DD-classifier performs well across a broad range of settings, and compares favorably with most existing nonparametric classifiers.