Skip to content

Latest commit

 

History

History
57 lines (33 loc) · 2.09 KB

RELEASE_NOTES.md

File metadata and controls

57 lines (33 loc) · 2.09 KB
Raw

Apriori - RELEASE NOTES

Version 2.0.1 (Apr. 13th 2019)

A minor release that introduces the following changes:

  • Updated Kotlin to version 1.3.21.
  • Updated Gradle configuration.

Version 2.0.0 (Aug. 30th 2018)

A major release, which introduces the following changes:

  • Migrated the library to use the Kotlin programming language instead of Java.

Version 1.3.0 (Dec. 13th 2017)

A feature release, which introduces the following changes:

  • The execute-method of the class Apriori now takes an Iterable as an argument. This allows to properly iterate transactions multiple times.
  • In accordance with the API specification, iterators are now expected to throw a NoSuchElementException instead of returning null when the end is reached.

Version 1.2.0 (Aug. 8th 2017)

A feature release, which introduces the following changes:

  • An enhanced API for sorting and filtering association rules and frequent item sets has been added
  • The class AssociationRule does now provide covers-methods, which allow to test, if the rule applies to a set of items
  • Added support for the conviction metric

Version 1.1.1 (Jul. 31th 2017)

A bugfix release, which fixes the following issues:

Version 1.1.0 (Jul. 30th 2017)

A feature release, which introduces the following changes:

  • Added the ability to use different tie-breaking strategies when sorting or filtering rule sets

Version 1.0.1 (Jul. 28th 2017)

A bugfix release, which fixes the following issues:

Version 1.0.0 (Jun. 16th 2017)

The first stable release of the library, which provides the following features:

  • Allows to find frequent item sets and (optionally) to generate association rules using the Apriori algorithm
  • The Apriori algorithm can be configured using the builder pattern
  • The algorithm can be configured to try to find a specific number of frequent item sets and/or association rules
  • Rule sets can be ordered and filtered by support, confidence, lift and leverage
  • Logging is provided by using the SLF4J logging facade