All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog and this project adheres to Semantic Versioning.
- Improve prediction performance
- Improve training performance and memory use
- New method
_get_raw_leaf_data()to access leaf internals
- Removed references to
SparseDataFrame(deprecated)
- Bug in regression code
- Better hyperplane tree plotting
- Bug in hyperplane trees (tried to access uninitialized field)
- Improved scikit-learn compatibility further
- Bug in
model.feature_importance()computation
- Experimental support for arbitrarily-oriented hyperplane splits rather than axis-perpendicular ones only
- Experimental support for sparse DataFrames and sparse matrices (scipy.sparse) for fitting and prediction
- Added
model.feature_importance()for feature selection - All models now compatible with scikit-learn models
- Lots of small changes here and there
- First release