Topology for preserving feature correlation in tabular synthetic data.

Arifeen, Murshedul; Petrovski, Andrei

Abstract

Tabular synthetic data generating models based on Generative Adversarial Network (GAN) show significant contributions to enhancing the performance of deep learning models by providing a sufficient amount of training data. However, the existing GAN-based models cannot preserve the feature correlations in synthetic data during the data synthesis process. Therefore, the synthetic data become unrealistic and creates a problem for certain applications like correlation-based feature weighting. In this short theoretical paper, we showed a promising approach based on the topology of datasets to preserve correlation in synthetic data. We formulated our hypothesis for preserving correlation in synthetic data and used persistent homology to show that the topological spaces of the original and synthetic data have dissimilarity in topological features, especially in 0th and 1st Homology groups. Finally, we concluded that minimizing the difference in topological features can make the synthetic data space locally homeomorphic to the original data space, and the synthetic data may preserve the feature correlation under homeomorphism conditions.

Citation

ARIFEEN, M. and PETROVSKI, A. 2022. Topology for preserving feature correlation in tabular synthetic data. In Proceedings of the 15th IEEE (Institute of Electrical and Electronics Engineers) International conference on security of information and networks 2022 (SINCONF 2022), 11-13 November 2022, Sousse, Tunisia. Piscataway: IEEE [online], pages 61-66. Available from: https://doi.org/10.1109/SIN56466.2022.9970505

• Presentation Conference Type: Conference Paper (published)
• Conference Name: 15th International conference on security of information and networks 2022 (SINCONF 2022)
• Start Date: Nov 11, 2022
• End Date: Nov 13, 2022
• Acceptance Date: Sep 25, 2022
• Online Publication Date: Nov 13, 2022
• Publication Date: Dec 16, 2022
• Deposit Date: Jan 9, 2023
• Publicly Available Date: Jan 9, 2023
• Publisher: Institute of Electrical and Electronics Engineers (IEEE)
• Peer Reviewed: Peer Reviewed
• Pages: 61-66
• Book Title: Proceedings of the 2022 15th IEEE International conference on security of information and networks 2022 (SINCONF 2022)
• ISBN: 9781665454650
• DOI: https://doi.org/10.1109/SIN56466.2022.9970505
• Keywords: Synthetic data; Correlation; GAN; Topology; Persistent homology
• Public URL: https://rgu-repository.worktribe.com/output/1853567