Department of Statistics Seminar
North Carolina State University

Presents

Susan Murphy

University of Michigan

A Prediction Interval for the Misclassification Rate
 
 
Abstract
 

The misclassification rate of a classifier is a non-smooth functions of the classifier.  The estimated rate suffers from bias due to over-fitting and the misclassification rate is a “minimized" quantity.  For both of these reasons the construction of measures of confidence such as estimates of variance and confidence/prediction intervals are challenging.   We discuss this problem and propose a method based on the use of a smooth upper bound combined with the bootstrap. This upper bound utilizes the surrogate loss that is used in the construction of the classifier.

Friday, October 31 2008
3:35 - 5:00 pm
321 Riddick Hall

Refreshments will be served in the common area of Riddick at 3:00 pm.

NOTE: No food or drink is allowed in any of the classrooms in Riddick Hall.