Area: Double;
The Area property returns curve area.
ROC curve area is an aggregated characteristic of classification quality that does not depend on the ratio of error values. The greater is the curve area value, the better is the classification model. This factor is often used for comparative analysis of multiple classification models.
Tolerance range: [0; 1].
The property use is given in the example for ISmGradientBoostedTree.ClassificationSummary.
See also: