Tools
Tools
Shinde, Aditya, Shinde, Swati, Virdee, Bal Singh and Khanna, Ashish (2026) Advancements in generative adversarial networks for synthetic data generation: A comprehensive analysis. In: Transforming Healthcare With AI And IoT: Intelligent Solutions for a Digital Future. Artificial Intelligence for Sustainable Engineering and Management . CRC Press (Routledge Taylor & Francis Group), Boca Raton (Florida, USA), pp. 160-180. ISBN 9781041111429; 9781003658504
This work investigates the use of Generative Adversarial Networks (GANs) for generating high-quality synthetic datasets with preserved statistical features of actual data. We implement a PyTorch version and train a GAN on tabular data, and we analyze the quality of generated synthetic samples using several statistical and visualization evaluation methods. The results show that well-trained GANs are capable of effectively modeling both the marginal distributions of a single feature and the intricate relationships amongst variables. Our work here describes an extensive evaluation protocol consisting of distribution comparisons, correlation analysis, dimensionality reduction visualizations, and statistical measures that collectively gauge the fidelity of the synthesized data. Our results indicate that GANs offer a viable method for synthetic data generation in scenarios where data privacy, augmentation, or simulation is necessary.
![[thumbnail of A4book chapterGAN for Synthetic Data Generation_new.pdf]](https://repository.londonmet.ac.uk/style/images/fileicons/text.png "A4book chapterGAN for Synthetic Data Generation_new.pdf")
Restricted to Repository staff only until 22 January 2027.
Available under License Creative Commons Attribution Non-commercial No Derivatives 4.0.
Download (1MB) | Request a copy
 |
View Item |