SSP'05 -- Proceedings

	IEEE/SP 13th workshop on Statistical Signal Processing July, 17-20, 2005 - Bordeaux - France

Information regarding the paper

Title

Minimum expected risk probability estimates for nonparametric neighborhood classifiers

Author(s)

Maya Gupta	University of Washington
Luca Cazzanti	University of Washington
Santosh Srivastava	University of Washington

Get the paper in PDF format

To obtain Acrobat Reader (version 5 minimum required) necessary to his read.

Abstract

We consider the problem of estimating class probabilities for a given feature vector using nonparametric neighborhood methods, such as k-nearest neighbors (k-NN). This paper's contribution is the application of minimum expected risk estimates for neighborhood learning methods, an analytic formula for the minimum expected risk estimate for weighted k-NN classifiers, and examples showing that the difference can be significant.