Testing the difference of the ROC curves in biexponential model

Rok, strany: 2008, 215 - 223

Receiver Operating Characteristics (ROC) curves are a useful instrument for evaluation of supervised classifiers performance. The classic form of ROC Receiver Operating Characteristics (ROC) curves are a useful instrument for evaluation of supervised classifiers performance. The classic form of ROC curve is applicable only to the two-stable classification. Neverthless, it covers a broad spectrum of applications (clinical diagnostics in medicine, computational linquistics, machine learning, data mining. etc.) In this contribution we will derive a test of equivalence of two ROC curves in biexponential ROC model for unpaired design of data. the result will be illustrated on real data comming from a sociological research.

ISO 690:

Betinec, M. 2008. Testing the difference of the ROC curves in biexponential model. In Tatra Mountains Mathematical Publications, vol. 39, no.1, pp. 215-223. 1210-3195.

APA:

Betinec, M. (2008). Testing the difference of the ROC curves in biexponential model. Tatra Mountains Mathematical Publications, 39(1), 215-223. 1210-3195.