Rejection sampling is filtering by rejecting samples. This technique can be used in synthetic generation as a post-processing so that samples with low fidelity can be rejected and will not be present in the final synthetic data set. An example is rejecting samples that do not meet deterministic business rules.