Are you evaluating synthetic data and benchmarking different synthesizers? Get help and file feedback here.
Check out the resources below for more information.
Synthetic Data Evaluation
Evaluate the quality and privacy of your synthetic data as compared to the real data.
- The SDV library offers native support for running quality and diagnostic checks on your data. See the docs for multi-table evaluation and single-table evaluation
- The SDMetrics library is our fully open-source, standalone library for running metrics, visualizing data, and generating reports . SDMetrics is model-agnostic, meaning you can use synthetic data created from any software (not just SDV).
Benchmarking Synthesizers
Compare different modeling algorithms. Our SDGym library allows you to compare synthesizers by applying them to a variety of different datasets. Compare the speed, memory usage, synthetic data quality, and more. This library natively supports all SDV synthesizers, but you can add in your own modeling algorithms too.