Comparative Study of the Estimation of the Area Under the ROC Curves
| Authors | |
|---|---|
| Year of publication | 2006 |
| Type | Article in Proceedings |
| Conference | Proceedings of the International Workshop on Spatio-Temporal Modelling (METMA3) |
| MU Faculty or unit | |
| Citation | |
| Field | Applied statistics, operation research |
| Keywords | ROC curve; parametric estimate; nonparametric estimate |
| Description | The aim of this contribution is to compare several methods of estimating the area under the reciever operating characteristic (ROC) curve for continous - scale diagnostic tests. The area under ROC curve is frequently used as a measure for the effectivness of diagnostic markers. Let AUC denote the area under this curve. Different estimates of the AUC are based on different approaches to estimating the ROC curve. In the present paper we investigate (i) parametric approach based on normal distribution or eliptically contoured distribution and (ii) nonparametric approach based on kernel smoothing. |
| Related projects: |