Skip to content

Enhance sccomp functions with unconstrained predictors#253

Merged
stemangiola merged 3 commits intomasterfrom
add-unconstrained-predictors
Feb 23, 2026
Merged

Enhance sccomp functions with unconstrained predictors#253
stemangiola merged 3 commits intomasterfrom
add-unconstrained-predictors

Conversation

@stemangiola
Copy link
Collaborator

  • Added output for unconstrained predictors in sccomp_predict and sccomp_calculate_residuals, providing users with direct access to linear scale predictions before softmax transformation.
  • Updated documentation to reflect new output columns, including unconstrained_mean, unconstrained_lower, and unconstrained_upper.
  • Implemented comprehensive tests to ensure the accuracy and availability of unconstrained predictors in both summary and draws modes.
  • Modified Stan model to save linear predictors, ensuring backward compatibility with older models.
  • Enhanced user understanding of model behavior through detailed documentation and testing.

- Added output for unconstrained predictors in `sccomp_predict` and `sccomp_calculate_residuals`, providing users with direct access to linear scale predictions before softmax transformation.
- Updated documentation to reflect new output columns, including `unconstrained_mean`, `unconstrained_lower`, and `unconstrained_upper`.
- Implemented comprehensive tests to ensure the accuracy and availability of unconstrained predictors in both summary and draws modes.
- Modified Stan model to save linear predictors, ensuring backward compatibility with older models.
- Enhanced user understanding of model behavior through detailed documentation and testing.
Updated version number and added details for version 2.1.27, including major enhancements and new output columns in sccomp functions.
@stemangiola stemangiola merged commit 3e6368e into master Feb 23, 2026
8 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant