qCBA v1.0.2 (Release date: 2025-04-03)
* Added possibility to redefine seed for qcbaIris(), qcbaIris2() and benchmarkQCBA
* Removed unnecessary library and require calls from package source
* the show method for the S4 class qCBARuleModel is now exported
 
qCBA v1.0.1 (Release date: 2024-09-08)
* Added print option for qCBARuleModel

qCBA v1.0 (Release date: 2024-08-26)
* Updates for CRAN release (disable long running examples)
* Minor bug fixes (supress incorrect warning in benchmarkQCBA())

qCBA v0.7 (Release date: 2024-07-31)
* Updated SBRL example to match new SBRL version 1.4

qCBA v0.6.1 (Release date: 2023-08-18)
Changes
* Bug fix: predict.qCBARuleModel when outputConfidenceScores=TRUE 

qCBA v0.6 (Release date: 2023-08-16)
==============
Changes
* Added new function benchmarkQCBA that bulk learns, QCBA postprocesses and evaluates 
rule models generated by arulesCBA  (CBA, CMAR, CPAR, PRM and FOIL2)
* Java-based QCBA implementation updated (added support for outputting rule length)
* Java backend built with Open JDK 17.
* Incorporates changes in coercions required by Matrix package (since 1.5-0)
* Updated documentation, added reference to the newly published article describing QCBA

qCBA v0.5.3 (Release date: 2023-07-07)
==============
Changes
* Added citation to the newly published article on QCBA

qCBA v0.5.2.1 (Release date: 2022-08-01)
==============
Changes
* Added support for converting new types of arulesCBA models (e.g., CPAR, FOIL) to qCBA datastructures

qCBA v0.5.1 (Release date: 2020-11-09)
==============
Changes

* Explainability enhancements: Predict function now supports outputing identifiers of rules used for classification and estimates of confidence scores (preview feature, not yet thoroughly tested).
* Sync with changes in arulesCBA version 1.2.0 with respect to the output of default rule
* Improved confidence estimation on the output of the prune function: In addition to standard measured computed by the arules library, new slots are available for each rule -- orderedConf and orderedSupp. Order-sensitive confidence  is computed only from instances reaching the given rule as a/(a+b), where a is the number of instances matching both the antecedent (available as slot orderedSupp) and consequent and b is the number of instances matching the antecedent, but not matching the consequent of the given rule.  

qCBA v0.4 (Release date: 2019-08-26)
==============
Changes

* Added support for sbrl models
* Fixed support of rCBA and arulesCBA models
* Fixed possible endless loop error in removal of redundant attributes

qCBA v0.3.1 (Release date: 2017-12-26)
==============
Changes

* Added examples to arulesCBAModel2CustomCBAModel and rcbaModel2CustomCBAModel functions