EvoSampling: A Granular Ball-based Evolutionary Hybrid Sampling with Knowledge Transfer for Imbalanced Learning

TL;DR

EvoSampling is a novel hybrid sampling method that uses evolutionary algorithms and granular ball techniques to improve data quality and class balance in imbalanced learning tasks, outperforming existing methods.

Contribution

It introduces a multi-granularity hybrid sampling approach combining genetic programming and granular ball-based undersampling with knowledge transfer, enhancing diversity and noise removal.

Findings

EvoSampling improves classification performance across 20 datasets.

It generates more diverse and high-quality minority class instances.

Knowledge transfer accelerates the evolutionary learning process.

Abstract

Class imbalance would lead to biased classifiers that favor the majority class and disadvantage the minority class. Unfortunately, from a practical perspective, the minority class is of importance in many real-life applications. Hybrid sampling methods address this by oversampling the minority class to increase the number of its instances, followed by undersampling to remove low-quality instances. However, most existing sampling methods face difficulties in generating diverse high-quality instances and often fail to remove noise or low-quality instances on a larger scale effectively. This paper therefore proposes an evolutionary multi-granularity hybrid sampling method, called EvoSampling. During the oversampling process, genetic programming (GP) is used with multi-task learning to effectively and efficiently generate diverse high-quality instances. During the undersampling process, we develop a granular ball-based undersampling method that removes noise in a multi-granular fashion, thereby enhancing data quality. Experiments on 20 imbalanced datasets demonstrate that EvoSampling effectively enhances the performance of various classification algorithms by providing better datasets than existing sampling methods. Besides, ablation studies further indicate that allowing knowledge transfer accelerates the GP's evolutionary learning process.