Some roc models with skewed distributions

Parametric ROC Curve is a mathematical model for assessing the performance of a binary classifier assuming that the data in both classes follow a known distribution. Construction of ROC model with skewed distributions has been the interest of researchers in the recent times. In this thesis we have focused on developing theoretical ROC model when both the distributions are skewed. We have used exponential and generalized exponential distributions to arrive at the new model. In this paper we have going to some important skewed distributions and their use in ROC modeling are discussed. We the skewed distribution and derive a new method of determining the AUC by utilizing the interval estimates of Scale and location parameters the distribution.