UncertaintyQuantification.jl icon indicating copy to clipboard operation
UncertaintyQuantification.jl copied to clipboard

Fix unstable tests

Open FriesischScott opened this issue 2 years ago • 2 comments

This should fix the unstable tests for empirical distributions and mappings to standard normal space.

Closes #148.

FriesischScott avatar Dec 23 '23 11:12 FriesischScott

Something seems very wrong here. It fails the test for samples drawn from a normal. They aren't even transformed. How is the pvalue this low?

FriesischScott avatar Dec 23 '23 11:12 FriesischScott

Are hypothesis test just useless? Even if I sample a standard normal, the test will return anything between 0 and 1.

FriesischScott avatar Dec 23 '23 15:12 FriesischScott

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Comparison is base (580a502) 98.27% compared to head (fd4b2d1) 98.27%.

Additional details and impacted files
@@           Coverage Diff           @@
##           master     #149   +/-   ##
=======================================
  Coverage   98.27%   98.27%           
=======================================
  Files          27       27           
  Lines        1045     1045           
=======================================
  Hits         1027     1027           
  Misses         18       18           

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

codecov[bot] avatar Jan 16 '24 15:01 codecov[bot]

While the hypothesis tests didn't work out, this at least improves the stability of the tests as far as I can tell.

FriesischScott avatar Jan 16 '24 16:01 FriesischScott