Generation of Synthetic Data for Sharing and Utilization in Healthcare Data

Ha Ye Jin Kang, Min Sam Ko, Kwang Sun Ryu

Abstract: The potential value of healthcare data in the context of artificial intelligence and data mining is widely acknowledged. However, there are significant limitations to the utilization and sharing of this data. In our work, we propose the production of synthetic data based on generative adversarial networks algorithms as a means of addressing this problem. The data employed in this study is the Cancer Public Staging Database (CPSD) in the National Cancer Data Centre. The experimental results demonstrate that the synthetic data exhibits a similar distribution to the original data. As a future work, it is necessary to compare the performance of the synthetic data based on predictive modelling to verify its objective effectiveness of the synthetic data.

External IDs:dblp:conf/dsc/KangKR24