Skip to main content

Table 5 Comparison of the performances of different strategies for class imbalance

From: CapsNet-SSP: multilane capsule network for predicting human saliva-secretory proteins

Strategies Accuracy Sensitivity Specificity Precision F-score MCC AUC
No strategy 0.853 0.796 0.909 0.897 0.843 0.710 0.916
Hybrid-based 0.868 0.857 0.879 0.875 0.866 0.736 0.939
Boosting-based 0.868 0.827 0.909 0.900 0.862 0.738 0.918
Bagging-based 0.888 0.847 0.929 0.922 0.884 0.779 0.948
  1. The threshold is set where the MCC reaches the maximum value