14
5

Exact Paired-Permutation Testing for Structured Test Statistics

Abstract

Significance testing -- especially the paired-permutation test -- has played a vital role in developing NLP systems to provide confidence that the difference in performance between two systems (i.e., the test statistic) is not due to luck. However, practitioners rely on Monte Carlo approximation to perform this test due to a lack of a suitable exact algorithm. In this paper, we provide an efficient exact algorithm for the paired-permutation test for a family of structured test statistics. Our algorithm runs in O(GN(logGN)(logN))\mathcal{O}(GN (\log GN )(\log N )) time where NN is the dataset size and GG is the range of the test statistic. We found that our exact algorithm was 1010x faster than the Monte Carlo approximation with 2000020000 samples on a common dataset.

View on arXiv
Comments on this paper