Close to Reality: Interpretable and Feasible Data Augmentation for Imbalanced Learning

TL;DR

This paper introduces DPG-da, an interpretable data augmentation framework for imbalanced learning that generates diverse, valid, and explainable samples by leveraging decision predicate graphs derived from trained models.

Contribution

The paper presents a novel, interpretable data augmentation method using decision predicate graphs to improve class imbalance handling.

Findings

DPG-da outperforms traditional over-sampling methods on benchmark datasets.

DPG-da ensures generated samples are diverse, valid, and interpretable.

The framework provides clear explanations for the augmented data.

Abstract

Many machine learning classification tasks involve imbalanced datasets, which are often subject to over-sampling techniques aimed at improving model performance. However, these techniques are prone to generating unrealistic or infeasible samples. Furthermore, they often function as black boxes, lacking interpretability in their procedures. This opacity makes it difficult to track their effectiveness and provide necessary adjustments, and they may ultimately fail to yield significant performance improvements. To bridge this gap, we introduce the Decision Predicate Graphs for Data Augmentation (DPG-da), a framework that extracts interpretable decision predicates from trained models to capture domain rules and enforce them during sample generation. This design ensures that over-sampled data remain diverse, constraint-satisfying, and interpretable. In experiments on synthetic and real-world benchmark datasets, DPG-da consistently improves classification performance over traditional over-sampling methods, while guaranteeing logical validity and offering clear, interpretable explanations of the over-sampled data.