Few-Shot Resampling for Scalable Statistically-Sound Data Mining

In three linesFewRS is a resampling method to assess statistical significance of data mining results. It reduces the number of resampled datasets needed (from thousands to few) while guaranteeing false discovery control. Tested on pattern mining and network analysis, it achieves up to 100× speedup.

Summary generated by Claude — human-verified