Apriori - Association Rule Induction / Frequent Item Set Mining

Download

Description

A program to find association rules and frequent item sets (also closed and maximal) with the apriori algorithm (Agrawal et al. 1993), which carries out a breadth first search on the subset lattice and determines the support of item sets by subset tests. This is a pretty fast implementation that uses a prefix tree to organize the counters for the item sets.

Full description of this program (included in the source package).

Attention: Please note that versions 3.x and 4.x differ from earlier versions in how they interpret the options -l, and -y. This is especially important for the option -l, which now means "do not load transactions into main memory", because using this option considerably reduces performance. Note also that since version 5.0 the default target type is sets rather than rules (in order to have a uniform default for all frequent item set mining programs).

Earlier versions of this program are incorporated in the data mining tool Clementine, available from SPSS (apriori version 1.8 in Clementine version 5.0, apriori version 2.7 in Clementine version 7.0, last version shipped to SPSS is 4.30).

A graphical user interface for this program (ARView), written in Java, is available here.

Another graphical user interface for this program, which is based on Gnome 2 and was developed by togaware, can be found here. Another graphical user interface, which is based on the wxWidgets and was developed by the STK++ team can be found here. (However, I cannot guarantee that these GUIs works with the latest version of the command line program made available here.)

Papers that describe some implementation aspects of this program:

• Recursion Pruning for the Apriori Algorithm
  Christian Borgelt.
  2nd Workshop of Frequent Item Set Mining Implementations (FIMI 2004, Brighton, UK).
  fimi_04.pdf (41 kb) fimi_04.ps.gz (29 kb) (2 pages)
• Efficient Implementations of Apriori and Eclat
  Christian Borgelt.
  Workshop of Frequent Item Set Mining Implementations (FIMI 2003, Melbourne, FL, USA).
  fimi_03.pdf (304 kb) fimi_03.ps.gz (197 kb) (9 pages)
• Induction of Association Rules: Apriori Implementation
  Christian Borgelt and Rudolf Kruse
  15th Conference on Computational Statistics (Compstat 2002, Berlin, Germany), 395-400
  Physica Verlag, Heidelberg, Germany 2002
  cstat_02.pdf (105 kb) cstat_02.ps.gz (91 kb) (6 pages)

Some other references:

• Fast Algorithms for Mining Association Rules
  R. Agrawal and R. Srikant
  Proc. 20th Int. Conf. on Very Large Databases (VLDB 1994, Santiago de Chile), 487-499
  Morgan Kaufmann, San Mateo, CA, USA 1994
• Fast Discovery of Association Rules
  R. Agrawal, H. Mannila, R. Srikant, H. Toivonen, and A. Verkamo
  In: U.M. Fayyad, G. Piatetsky-Shapiro, P. Smyth, and R. Uthurusamy, eds.
  Advances in Knowledge Discovery and Data Mining, 307-328
  AAAI Press / MIT Press, Cambridge, CA, USA 1996

More information about frequent item set mining, implementations of other algorithms as well as test data sets can be found at the Frequent Itemset Mining Implementations Repository.