Random subsampling correction

meelgroup / barbarik

The first efficient procedure to test whether a sampler is uniform

MIT License

13 stars 1 forks source link

Random subsampling correction #11

Closed yashpote closed 3 years ago

yashpote commented 3 years ago

Given N samples when K (N>K) samples are required, Barbarik should randomly choose the K from the set of N. However, the implementation currently does it in a way that may introduce bias. This pull request takes care of that issue.

msoos commented 3 years ago

Looks cool! Would it be okay if I review this on Sunday? Cheers,

Mate

msoos commented 3 years ago

Just some minor feedback: we need to have tests. Currently, we don't have tests and this makes it VERY hard to reason about this program. It'd be nice to have some tests...