Imbalanced Classification via a Tabular Translation GAN

TL;DR

This paper introduces a GAN-based method for imbalanced binary classification on tabular data, generating synthetic minority samples to improve classifier performance, especially in severe class imbalance scenarios.

Contribution

It proposes a novel translation GAN with regularization and sample selection to enhance minority class modeling in imbalanced tabular datasets.

Findings

Improves average precision over re-weighting and oversampling methods.

Effective in severe class imbalance scenarios.

Enhances minority class representation with synthetic samples.

Abstract

When presented with a binary classification problem where the data exhibits severe class imbalance, most standard predictive methods may fail to accurately model the minority class. We present a model based on Generative Adversarial Networks which uses additional regularization losses to map majority samples to corresponding synthetic minority samples. This translation mechanism encourages the synthesized samples to be close to the class boundary. Furthermore, we explore a selection criterion to retain the most useful of the synthesized samples. Experimental results using several downstream classifiers on a variety of tabular class-imbalanced datasets show that the proposed method improves average precision when compared to alternative re-weighting and oversampling techniques.