发表机构
Department of Informatics and Telecommunications, University of Ioannina; Industrial Systems Institute, Athena Research Center; Archimedes Unit, Athena Research Center; Department of Computer Engineering and Informatics, University of Patras; Department of Mechanical Engineering and Aeronautics, University of Patras（约阿尼纳大学信息与电信系; 雅典娜研究中心工业系统研究所; 雅典娜研究中心阿基米德单位; 帕特雷大学计算机工程与信息学系; 帕特雷大学机械工程与航空系）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
研究针对表格分类中过采样会扭曲特征重要性排名的问题，提出K-IPO框架，通过“生成然后选择”保留原始数据特征重要性排名，经实验评估，该方法在多方面表现出色，提升了预测性能且计算开销合理。
AI 中文摘要
过采样广泛用于解决表格分类中的类不平衡问题，但现有方法会扭曲模型解释背后的特征重要性排名。虽有研究量化此扭曲，但无人主动预防。本文引入肯德尔约束重要性保持过采样（K-IPO），这是一种与生成器无关的“生成然后选择”框架，在增强过程中保留原始数据的特征重要性排名。K-IPO迭代生成少数类候选样本，仅当包含它们能与参考排名保持用户定义的最小肯德尔tau（τ）相关性时才接受。可对最高排名特征应用更严格约束。我们用三个分类器和多种解释方法在20个不平衡二元分类数据集上评估了K-IPO。多数情况下，K-IPO在特征重要性保持、解释一致性和类可分性方面取得最佳或并列最佳结果，还在保持有竞争力的计算开销时普遍提高了预测性能。
英文摘要
Oversampling is widely used to address class imbalance in tabular classification, but existing methods can distort the feature importance ranking underlying model explanations. Although recent studies have quantified this distortion by comparing real and synthetic data, none have actively sought to prevent it. In this paper, we introduce Kendall-constrained Importance-Preserving Oversampling (K-IPO), a generator-agnostic, "generate-then-select" framework that preserves the original data's feature importance ranking during augmentation. K-IPO iteratively generates minority-class candidates and accepts them only if their inclusion maintains a user-defined minimum Kendall's tau (τ) correlation with the reference ranking. Optionally, stricter constraints can be applied to the highest-ranked features. We evaluated K-IPO on 20 imbalanced binary classification datasets using three classifiers and multiple explanation methods. In most cases, K-IPO achieved the best or tied-best results in feature importance preservation, explanation consistency, and class separability. It also generally improved predictive performance while maintaining competitive computational overhead.
Comments24 pages, 5 figures, 9 tables
