v2.1.15: First public RuleKit 2 release

Removed RapidMiner dependency

v1.7.19: Contrast sets induction working properly again

Some logging added during contrast sets induction.

v1.7.14: Added `mincov_new` adjustment at the end of growing (results may change).

Added mincov_new adjustment at the end of growing. Small fixes making the algorithm insensitive to the value ordering of nominal attributes.

v1.7.11: Removed mincov adjustment.

Removed strange mincov adjustment (results may change).

v1.7.10: Several fixes in handling nominal attributes in the approximate induction.

Several fixes in handling nominal attributes in approximate induction.

v1.7.9: Values of nominal attributes analyzed in alphabetical order

Nominal attribute values investigated in the alphabetical order.

v1.7.8: Classes analyzed in alphabetical order

Classes analyzed in alphabetic order.

v1.7.7 Small bugfix

Fixed very rare bug of model being dependent on the order in the nominal attribute mapping.

v1.7.6: Better error reporting when parsing expert knowledge

More detailed reporting of the rule parsing errors.

v1.7.5: Small change in the expert induction parameters

Update in preferredAttributesPerRule/preferredConditionsPerRule parameters: the value of 0 interpreted as the infinity.