kappa-rho-conserv decision tree

[handwerkerd]

I've discussed with a bunch of people the idea of making a conservative decision tree that will not remove all noise, but is very unlikely to remove signal someone cares about.

At first glance, this is sometimes retaining the high variance linear drift component, so maybe we might want this to be a bit more aggressive. Still, if the goal is to be conservative, then that's not necessary a problem. I'm releasing this as a draft so that people can test this on their data to see if it performs as expected.

Changes proposed in this pull request:

• New decision tree is tedana/resources/decision_trees/kapparho_conserv.json
  • This is based on the minimal decision tree expect I got rid of all other metrics besides kappa & rho
  • The basic logic is to just use the kappa & rho elbows, but retain components that are >0.95& kappa elbow and <1.05* rho elbow.
• tedana/selection/component_selector.py is also updated so that this decision tree would be one of the default options.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 62.34%. Comparing base (dd24fef) to head (8cd99ec).
⚠️ Report is 7 commits behind head on main.

❗ There is a different number of reports uploaded between BASE (dd24fef) and HEAD (8cd99ec). Click for more details.

HEAD has 1 upload less than BASE

Flag	BASE (dd24fef)	HEAD (8cd99ec)
	7	6

Additional details and impacted files

@@             Coverage Diff             @@
##             main    #1449       +/-   ##
===========================================
- Coverage   90.41%   62.34%   -28.08%     
===========================================
  Files          30       30               
  Lines        4906     4903        -3     
  Branches      901      898        -3     
===========================================
- Hits         4436     3057     -1379     
- Misses        316     1695     +1379     
+ Partials      154      151        -3     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.