How to evaluate the quality of synthetic data: fidelity, utility, and privacy
Synthetic data is only useful if you can prove it's faithful, useful and private.
By Faris Haddad · Published Dec 2022 on AWS Blog - Why it matters
- A framework for measuring synthetic data on fidelity, utility and privacy, for regulated firms that can't freely use real data.
- What you'll get
- An evaluation method your risk team will accept.
- Who should read it
- Data scientists, privacy officers
This page is a short guide. The full article is published on AWS Blog.