|
|
 |
| |
|
|
|
Statistical Analysis of the BIESEP ROC Curve |
|
|
|
PP: 57-66 |
|
|
|
doi:10.18576/jsapl/070202
|
|
|
|
Author(s) |
|
|
|
Ahmad Flaih,
Chary Akmyradov,
Jose Guardiola,
Hassan Elsalloukh,
|
|
|
|
Abstract |
|
|
| In this work, we consider a case study for the Bi-Epsilon Skew Exponential Power (BIESEP) ROC curve proposed by Flaih et al. [3]. This model is a generalization of the Epsilon Skew bi-normal ROC curve proposed by Mashtare Jr. and Huston [10]. Elsalloukh [1,2] provided the Epsilon Skew Exponential Power (ESEP) which is less sensitive to outliers. The ESEP family can be adopted to cope with skewness and kurtosis of a data set. The ESEP model provides an appropriate choice to increase the robustness of data analysis. We develop the Epsilon Skew bi-normal ROC curve based on the outcomes of the diagnostic test that is distributed according to the ESEP distribution. More specifically, we derive the BIESEP ROC parameters and the Area Under the Curve (AUC). Also, we consider the parameter estimation of BIESEP ROC curve and the AUC of a diagnostic test. We employ the BIESEP ROC curve to analyze a real dataset. |
|
|
|
|
|
|
|
