Generating ROC curves for artificial neural networks.

Abstract

Receiver operating characteristic (ROC) analysis is an established method of measuring diagnostic performance in medical imaging studies. Traditionally, artificial neural networks (ANN's) have been applied as a classifier to find one "best" detection rate. Recently researchers have begun to report ROC curve results for ANN classifiers. The current standard method of generating ROC curves for an ANN is to vary the output node threshold for classification. In this work, we propose a different technique for generating ROC curves for a two-class ANN classifier. We show that this new technique generates better ROC curves in the sense of having greater area under the ROC curve (AUC), and in the sense of being composed of a better distribution of operating points.