关键词: 不均衡数据集 过取样 欠取样 代价敏感学习

Abstract: The majority of machine learning algorithms previously designed usually assume that their training sets are well-balanced,but data in real-world is usually imbalanced. The tradition machine learning algorithms on balanced data sets have bad performance wh

Key words: Imbalance data set, Over-sampling,Under-sampling Cost-sensitive learning