MargCTGAN: A "Marginally'' Better CTGAN for the Low Sample Regime

TL;DR

This paper evaluates synthetic tabular data generators in low sample regimes, identifies CTGAN's limitations, and proposes MargCTGAN, which improves statistical properties and utility in such scenarios.

Contribution

Introduces MargCTGAN, a novel method enhancing CTGAN by feature matching of de-correlated marginals for better low sample performance.

Findings

MargCTGAN outperforms CTGAN in low sample settings.

Statistical properties of synthetic data are improved with MargCTGAN.

Downstream utility is consistently better with MargCTGAN in low sample regimes.

Abstract

The potential of realistic and useful synthetic data is significant. However, current evaluation methods for synthetic tabular data generation predominantly focus on downstream task usefulness, often neglecting the importance of statistical properties. This oversight becomes particularly prominent in low sample scenarios, accompanied by a swift deterioration of these statistical measures. In this paper, we address this issue by conducting an evaluation of three state-of-the-art synthetic tabular data generators based on their marginal distribution, column-pair correlation, joint distribution and downstream task utility performance across high to low sample regimes. The popular CTGAN model shows strong utility, but underperforms in low sample settings in terms of utility. To overcome this limitation, we propose MargCTGAN that adds feature matching of de-correlated marginals, which results in a consistent improvement in downstream utility as well as statistical properties of the synthetic data.