NEWS
SensIAT 0.3.0.9000
- Added generalized fitting support across marginal mean and within-group workflows, including flexible link/loss handling and GLM outcome-model support.
- Added new simulation and resampling capabilities, including intensity/outcome simulator factories and a parametric bootstrap workflow.
- Improved performance and numerical stability for generalized fitting and influence-term computation, including caching and additional integration-method support.
- Expanded tests, CI checks, and documentation, alongside multiple bug fixes discovered during broader generalized-model and bootstrap development.
SensIAT 0.3.0 (2025-09-05)
- Conformed most names to standard of starting with verbs.
- Documentation fixes and updates.
- Removed unnecessarily exported functions.
- Updated single index model with norm 1 coefficients to allow for multiple iterations.
- Various bug fixes.
- Source code cleanup.
SensIAT 0.2.0 (2025-08-18)
- New options for fitting single index outcome models:
MAVE (Minimum Average Variance Estimation)
- Fixed bandwidth, where the bandwidth is held constant and all coefficients are allowed to vary freely.
SensIAT 0.1.1 (2024-11-17)
- Bug fix for compiling on Debian and other systems without C++20.
SensIAT 0.1.0 (2024-11-01)
- Initial CRAN submission.
- Beta Release.
- Includes support for
'dnorm' and 'K2_biweight'(quartic) kernels.