Release 0.0.9

Fix for the Regressor.
With sparse feedback as the error nears 0 the estimator would not be updated.

Release 0.0.8

• bitwise algorithm
• weighted regressor
• speed improvements

Release 0.0.7

Changes for pyTsetlini

Classic regressor

Release 0.0.6 includes implementation of classic regressor Tsetlin Machine.
Based on bitwise CAIR implementation https://github.com/cair/pyTsetlinMachine

https://arxiv.org/abs/1905.04206

Release 0.0.5

TM is built upon number of clauses which properly reflects meaning of the number_of_pos_neg_clauses_per_label parameter. Previously, TM was twice the size.
Clause feedback calculation was redone to exactly reflect original CAIR implementation.

Faster classic Classifier

Merge pull request #34 from WojciechMigda/GH33-data-processing-diet

Calculate clause output and train automata only using data for the two selected labels

Partial fit works as expected

Merge pull request #21 from WojciechMigda/GH20-partial-fit-example

Add example with partial fit which restart using deserialized state

Release for PyTsetlini

v0.0.2

Add some extra compiler features/STL checks

Initial release

v0.0.1

Update deprecated setup.py