数据不平衡分类：基于集成采样和Adaboost的改进算法

The comparison algorithm used undersampling algorithms RUS, ENN, and CNN, oversampling algorithms SMOTE, Borderline SMOTE, ADASYN, and Near Miss, combined oversampling and undersampling algorithm SMOTE-ENN, as well as ensemble algorithms SMOTEBoost and RUSBoost. Except for ensemble algorithms, which can directly perform classification, other sampling algorithms use Adaboost as the classifier after resampling the training set. Table 2 shows the AUC classification metrics of the proposed method and the comparison methods on 10 datasets, and Table 3 shows the G_mean classification metrics of the proposed method and the comparison methods on 10 datasets.