What measurement indicates the ability of a model to predict higher scores for positive examples compared to negative examples?

The Area Under the ROC Curve (AUC) is a robust metric that quantifies a model's ability to differentiate between positive and negative classes. It specifically measures the model's performance across various classification thresholds, illustrating how well it ranks positive examples higher than negative ones.

When calculating AUC, the ROC (Receiver Operating Characteristic) curve is plotted based on the true positive rate (sensitivity) against the false positive rate. AUC provides a single scalar value between 0 and 1, where an AUC of 0.5 indicates no discrimination (the model cannot distinguish between classes), and an AUC of 1.0 signifies perfect classification.

This metric is particularly useful in scenarios where there is a strong class imbalance, as it reflects the true positive rate while accounting for the false positive rate across varying thresholds, rather than simply focusing on classifications made at a single decision threshold. This property of AUC effectively captures how well the model ranks higher scores for positive examples in comparison to negative examples, providing insight into its predictive capabilities.