Fix oversized LightGBM default for CATE scoring #1037

[su-jin1425]

Proposed changes

Reduce the default num_leaves for the LightGBM learners used by CATE scoring from 64 to 31.

These learners are fit per fold and per treatment arm, so the previous default was unnecessarily large for the smaller training subsets and could result in unnecessary computation.

This updates the defaults in:

• compute_dr_pseudo_outcomes
• dr_score
• plug_in_t_score

Also adds/strengthens tests covering the default learner paths.

Fixes #1037.

Before and After

Test

$ErrorActionPreference="Stop"; $Base="$env:TEMP\causalml-1037-check"; if(Test-Path $Base){Remove-Item $Base -Recurse -Force}; git clone -q https://github.com/uber/causalml.git "$Base\before"; git clone -q https://github.com/su-jin1425/causalml.git "$Base\after"; Set-Location "$Base\before"; git checkout -q 9a2327811406240925e9b8c0ce16734a04da52d7; Write-Host "`n========== BEFORE #1037 ==========" -ForegroundColor Cyan; Select-String -Path "causalml\metrics\cate_scoring.py" -Pattern "num_leaves"; Set-Location "$Base\after"; git checkout -q f53395ac8085ae44f46bbf5f3a826b23ba536b2e; Write-Host "`n========== AFTER #1037 ==========" -ForegroundColor Green; Select-String -Path "causalml\metrics\cate_scoring.py" -Pattern "num_leaves"; Write-Host "`n========== RESULT ==========" -ForegroundColor Yellow; Write-Host "BEFORE: 64 leaves"; Write-Host "AFTER : 31 leaves"; Write-Host "Issue #1037 fix verified."

Types of changes

• Bugfix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Documentation Update (if none of the other choices apply)

Checklist

• I have read the CONTRIBUTING doc
• I have signed the CLA
• Lint and unit tests pass locally with my changes
• I have added tests that prove my fix is effective or that my feature works
• I have added necessary documentation (if appropriate)
• Any dependent changes have been merged and published in downstream modules

Further comments

The change keeps the existing LightGBM learner configuration unchanged apart from num_leaves, reducing the model size for the per-fold, per-arm fitting pattern without changing the public API.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.