mschauer
WIP: independence tests for categorical variables
implemented permutation-based independence tests for categorical variables based on mutual information and conditional mutual information
open questions:
- can type information be used more elegantly/efficiently?
- is the conditional mutual information test implemented correctly? do we need more tests to be sure?
- are there more efficient entropy estimators that we should implement?
Codecov Report
Merging #11 into master will increase coverage by
0.95%. The diff coverage is100%.
@@ Coverage Diff @@
## master #11 +/- ##
==========================================
+ Coverage 83.99% 84.95% +0.95%
==========================================
Files 7 7
Lines 606 638 +32
==========================================
+ Hits 509 542 +33
+ Misses 97 96 -1
| Impacted Files | Coverage Δ | |
|---|---|---|
| src/pc.jl | 89.18% <100%> (+0.79%) |
:arrow_up: |
| src/klentropy.jl | 89.21% <100%> (+3.31%) |
:arrow_up: |
| src/skeleton.jl | 89.55% <100%> (+1.62%) |
:arrow_up: |
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact),ø = not affected,? = missing dataPowered by Codecov. Last update f354ca8...b79fa8e. Read the comment docs.
Coverage increased (+1.0%) to 84.953% when pulling b79fa8e79fd710ee627518252dbb5e5e27781e5b on categorical_tests into f354ca8ce9052293023f7ccefda299bf946dd977 on master.
Do you think you can make this work without my input? Please merge once you think this is fine.
@mschauer I finally have some time this month to do more work on this, will let you know once I have some sensible ready.
This is great news. There has been also some interest for this package in the last weeks, so this would be a good time.