search

derrynknife / SurPyval

A Python package for survival analysis. The most flexible survival analysis package available. SurPyval can work with arbitrary combinations of observed, censored, and truncated data. SurPyval can also fit distributions with 'offsets' with ease, for example the three parameter Weibull distribution.
https://surpyval.readthedocs.io/en/latest/index.html
MIT License
47 stars 5 forks source link

[Joss review] Test tolerance and sample sizes #17

Closed CamDavidsonPilon closed 3 years ago

CamDavidsonPilon commented 3 years ago

I'd like to see tighter tolerances in the tests, and tests with smaller samples sizes. Currently the TOL is set to 0.2 - this is quite a large margin, practically useless in some cases. The minimum sample size is 5000 - that's medium-sized data, but how well do the algorithms work for small size data (sub 100, for example).

https://github.com/openjournals/joss-reviews/issues/3484

derrynknife commented 3 years ago

Fair, I changed it to 20% because the git workflows seemed to not converge as well as my local machine. I've changed it back to much finer tolerances and split the convergence into two tests. A large scale convergence and a small sample convergence.

commit d4df11ff7c13aa372ee843d6585751282706de98