Area under the curve and PR curve

The output first reminds users of the existence of the ties in the test variable, which is not uncommon when you analyze a real data set. The Case Processing Summary table summarizes the positive and the negative counts. The counts are split as males and females.

Figure 1. Case Processing Summary table
Case Processing Summary table

Both the ROC and PR curves are plotted by gender. By reviewing the charts, it is discovered that the two ROC curves are pretty close to each other, while males and females have difference in the precision and recall estimates.

Figure 2. ROC Curves chart
ROC Curves chart
Figure 3. Precision-Recall Curves chart
Precision-Recall Curves chart

The point estimates of the area under the ROC curve are 0.566 and 0.575 for males and females, respectively.

Figure 4. Area Under the ROC Curve table
Area Under the ROC Curve table

In the Independent-Group Area Difference Under the ROC Curve table, the estimated two-sided p-value is 0.685, based on which we fail to reject the null hypothesis that the true area difference between males and females is zero.

Figure 5. Independent-Group Area Difference Under the ROC Curve table
Independent-Group Area Difference Under the ROC Curve table