Table 5

Results of the evaluation metrics for each under-sampling method

Under-sampling methodAUPRCMAUCG-meanMMCC
RUS0.6310.7830.5320.243
Near Miss0.6630.9220.7600.564
OSS0.6710.9120.7770.571
NCR0.8230.9740.9110.814

Note(s): AUPRC, Area Under the Precision-Recall Curve; MAUC, Mean Area Under the Curve; MMCC, Mean Mathews Correlation Coefficient; RUS, Random Under-sampler; OSS, One-sided Selection; NCR, Neighborhood Cleaning Rule. For each of the evaluation metrics, a higher score (closer to 1) represents a better prediction performance

or Create an Account

Close subscription notice
Close access options