Scikit-psl

Added

- Significantly extended the configuration capabilities with predefined features to limit the PSLs searchspace

Changed

- PSL global loss defaults to sum(cascade)
- rewrote/extracted expected entropy calculation

Fixed

- PSL inspect is now more robust

Fixed

- PSL classifier optimization regarding global loss was incorrect

Added

- _ClassifierAtK
- Sigmoid calibration additional to isotonic
- PSL classifier
- Make optimization loss configurable
- Small `searchspace_analyisis(·)` function makes lookahead choice more informed

Fixed

- Fixed lookahead search space and considering global loss for model-sequence evaluation

Changed

- Updated dependencies and added black
- Moved Binarizer to different module
- Moved PSL hyperparameters to constructor

Fixed

- _ClassifierAtK
- Expected entropy for stage 0 now also calculated wrt. base 2
- Data with only 0 or 1 is now also interpret as binary data

Fixed

- Small import error

Added

- Add brute force threshold optimization method to find the global optimum, bisect optimizer remains default method

Changed

- Restructured source files