NEWS
fastLogisticRegressionWrap 1.2.2
- Made
assert_binary_vector_then_cast_to_numeric() faster, improving regression performance.
fastLogisticRegressionWrap 1.2.1
- Added a function that returns desired asymmetric-cost models fitted out of sample.
fastLogisticRegressionWrap 1.2.0 (2023-08-08)
- Added a function that returns desired asymmetric-cost models.
- Added alphabetical ordering of coefficients to the inference summary function.
- Sped up
confusion_results() with a custom C++ table routine and added an option to skip argument checks.
- Fixed inference-disabled operation.
- Fixed
proportions_scaled_by_column = TRUE in general_confusion_results().
fastLogisticRegressionWrap 1.1.0
- Added support for custom matrix-multiplication and square-root-diagonal-of-inverse functions, enabling GPU-backed computation (see the README for an example).
- Added conjugate-gradient inference for a single coefficient, which can be faster than inference for all coefficients in some settings.
fastLogisticRegressionWrap 1.0.1 (2023-07-08)
- Removed the OpenMP dependency for CRAN compatibility.
fastLogisticRegressionWrap 1.0.0 (2023-07-07)