28518378926
85%
main: 85%

Ran 01 Jul 2026 12:49PM UTC

Jobs 1

Files 31

Run time 1min

Badge

Embed ▾

Committed 01 Jul 2026 12:43PM UTC coverage: 71.914% (+0.3%) from 71.578%

Build # 28518378926

Build Type

push

github

Committed by

NilsWinter

Commit Message

Add scikit-learn equivalence tests for edge selection and CPM models

Independently validate the toolbox's vectorised torch/CUDA implementation
against scipy/numpy/scikit-learn reference implementations:

- Pearson, partial (semipartial) and Spearman edge statistics match the
  textbook definitions (effect sizes to 1e-6; p-values to 2e-3, and identical
  selected-edge masks at p<0.05).
- The batched torch OLS solvers for all four CPM model variants
  (connectome/covariates/full/residuals) match sklearn LinearRegression when
  given identical edges (predictions to 2e-3).
- The full cross-validated pipeline reproduces an independent sklearn pipeline
  and exhibits the raw > partial > residualised ~ true confound-inflation
  ordering.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

Coverage Stats

2138 of 2973 relevant lines covered (71.91%)

0.72 hits per line

Jobs

ID	Job ID	Ran	Files	Coverage
1	28518378926.1	01 Jul 2026 12:49PM UTC	31	71.91	GitHub Action Run

wwu-mmll / confound_corrected_cpm / 28518378926
85%
main: 85%

README BADGES
x

Markdown

Textile

RDoc

HTML

Rst

Jobs

Source Files on build 28518378926

wwu-mmll / confound_corrected_cpm / 28518378926 85% main: 85%

README BADGES x

Markdown

Textile

RDoc

HTML

Rst

Jobs

Source Files on build 28518378926

wwu-mmll / confound_corrected_cpm / 28518378926
85%
main: 85%

README BADGES
x