View Image

A program to find frequent item sets (also closed and maximal) with the relim algorithm (recursive elimination), which is inspired by the FP-growth algorithm, but does its work without prefix trees or any other complicated data structures. The main strength of this algorithm is not its speed (although it is not slow, but even outperforms apriori and eclat on some data sets), but the simplicity of its structure. Basically all the work is done in one recursive function of fairly few lines of code.

Papers that describes the algorithm underlying this program (that is, the program relim, see above):

Keeping Things Simple: Finding Frequent Item Sets by Recursive Elimination
Christian Borgelt.
Workshop Open Source Data Mining Software (OSDM'05, Chicago, IL), 66-70.
ACM Press, New York, NY, USA 2005
relim.pdf (166 kb) relim.ps.gz (122 kb) (5 pages)
Simple Algorithms for Frequent Item Set Mining
Christian Borgelt
Advances in Machine Learning II (Studies in Computational Intelligence 263), ???-??? (to appear)
J. Koronacki, Z.W. Raz, S.T. Wierzchon, and J.K. Kacprzyk (eds.)
Springer-Verlag, Berlin, Germany 2010
simple.pdf (416 kb) simple.ps.gz (261 kb) (18 pages)
An extension of the algorithm to mine fuzzy frequent item sets (which is available as part of this program, but was originally implemented as a separate program relx) has been developed in:

Mining Fuzzy Frequent Item Sets
Xiaomeng Wang, Christian Borgelt, and Rudolf Kruse
Proc. 11th Int. Fuzzy Systems Association World Congress (IFSA'05, Beijing, China), 528-533.
Tsinghua University Press and Springer-Verlag, Beijing, China, and Heidelberg, Germany 2005.
ifsa_05.pdf (107 kb) ifsa_05.ps.gz (70 kb) (6 pages)
Fuzzy Frequent Pattern Discovery Based on Recursive Elimination
Xiaomeng Wang, Christian Borgelt, and Rudolf Kruse
Proc. 4th IEEE Int. Conf. on Machine Learning and Applications (ICMLA'05, Los Angeles, CA), 391-396.
IEEE Press, Piscataway, NJ, USA 2005
icmla_05.pdf (159 kb) icmla_05.ps.gz (140 kb) (6 pages)