Fig. 4.

Receiver Operating Characteristic curve for the proposed stacked ensemble model. This plot measures the classification performance at various decision threshold values by showing the model’s ability at distinguishing both classes. As the blue curve gets closer to the upper and left-hand boundaries, the model shows a smaller Type I error. The red line illustrates the case in which the model is unable to tell one class from the other, being no better than classifying by chance. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)