Changelog

Source: NEWS.md

rCISSVAE v1.0.0

NOTE

• Major changes made to ensure correct handling of categorical variables
• Cross-entropy loss now used for categorical variables
• Please use update_cissvae_env() to upgrade your python ciss-vae package ## Updates
• Added categorical_column_map to fix the preparation of valdata for caegorical columns
• Added update_cissvae_env to update the cissvae environment when needed
• Added impute_result class and save/load methods for impute_result class
• save_cissvae_model() and load_cissvae_model() can now save and load state_dicts as well as full models. State_dict is preferred

rCISSVAE v0.0.5

CRAN release: 2026-03-03

Updates

• performance_by_cluster() imputation error calculation now matches python implementation
• changed all columns_ignore parameters to cols_ignore. Keeping columns_ignore as an alias for continuity.

New additions

• save_cissvae_model() will save CISSVAE models to disk
• load_cissvae_model() loads a saved CISSVAE model from disk
• impute_with_cissvae() accepts a model and R data.frame with missingness and uses the model to impute the data

rCISSVAE v0.0.4

CRAN release: 2026-01-23

Updates

• added missingness heatmap function cluster_heatmap() and associated vignette
• Updated the examples to be ‘donttest{}’ not ‘dontrun{}’
• Added checks for reticulate to functions requiring reticulate
• Added tutorial for loading and imputing with saved model

rCISSVAE v0.0.3

Updates

• binary_feature_mask is now correctly recognized by both autotune_cissvae() and run_cissvae()
• Added tests for making sure binary_feature_mask is correctly recognized

rCISSVAE v0.0.2.0000

Updating for CRAN readiness

Updates

• Improved Vignettes
  • Added extdata files for sample results
• Added documentation for undcumented things

rCISSVAE v0.0.1.0000

Initial Package Version

Updates

• gtsummary-like cluster_summary function