Average Precision for Partitioning Scenario N1: Average Precision (AP), also equivalent to area under the precision–recall curve (AUPRC or PRAUC) with no interpolation, ranges from 0 to 1. For more information, see sklearn's documentation. The goal is to maximize AP. In the case of multilabel classification, this metric will be calculated as a macro-average, with the final metric being an unweighted mean of AP values calculated for each label separately.

Average Precision for Partitioning Scenario N2: Average Precision (AP), also equivalent to area under the precision–recall curve (AUPRC or PRAUC) with no interpolation, ranges from 0 to 1. For more information, see sklearn's documentation. The goal is to maximize AP. In the case of multilabel classification, this metric will be calculated as a macro-average, with the final metric being an unweighted mean of AP values calculated for each label separately.

Average Precision for Partitioning Scenario N3: Average Precision (AP), also equivalent to area under the precision–recall curve (AUPRC or PRAUC) with no interpolation, ranges from 0 to 1. For more information, see sklearn's documentation. The goal is to maximize AP. In the case of multilabel classification, this metric will be calculated as a macro-average, with the final metric being an unweighted mean of AP values calculated for each label separately.