Applications of Frequent Pattern Mining - Frequent Pattern Mining - page 452

13 Resources for the Practitioner

Since frequent pattern mining methods are used frequently for different applications,
it is helpful to use off-the-shelf software for frequent pattern mining in many of these
applications. In addition, software for some of the applications discussed in this
chapter are also available.
A general Web site containing pointers to different resources on frequent pat-
tern mining is the KDD Nuggets site [ 146 ]. This can be considered an excellent
meta-repository containing pointers to frequent pattern mining software. The Weka
repository [ 154 ] contains many implementations of different data mining algorithms
including frequent pattern mining. In terms of specific implementations of different
algorithms, an implementation by Bart Goethals of some of the more well known
frequent pattern mining algorithms such as Apriori , Eclat , DIC and FP-Growth may
be found at [ 155 ]. A fast implementation of Apriori which uses prefix trees may be
found in [ 149 ]. This site also contains a significant amount of software for other
algorithms such as Eclat , FP-Growth , closed pattern mining, and maximal pattern
mining. Methods for fault tolerant and sequence mining are also covered by this
software collection. A good set of implementation of the FP-Growth family of al-
gorithms may also be found in [ 157 ]. The ARtool [ 156 ] is an open-source software
which is available under the GNU public license and is a collection of software and
tools for performing association analysis in market basket data sets. A well known
repository for different implementations of frequent pattern mining algorithms is
the FIMI repository [ 145 ]. This is an open source repository containing many ef-
ficient implementations of frequent pattern mining algorithms. In addition, a free
R-software package arules , which can perform frequent pattern mining of different
kinds is available in [ 150 ]. A significant amount of software is also available for rule-
based classification. The CBA system for classification with frequent pattern mining
[ 87 ] is available as an implementation at [ 147 ]. The rule-based system known as
RIPPER is available at [ 148 ].
In terms of commercial software, the IBM SPSS system contains software for dis-
covering frequent patterns and association rules from transaction and market basket
data. While this software is a general data mining platform, it also contains sig-
nificant parts which are tailored to market basket analysis. Oracle Data Mining is
a general purpose data mining tool, which also provides association mining capa-
bilities [ 158 ]. SAS provides an Enterprise Miner , which provides the capability to
mine both associations and sequential patterns [ 159 ]. The SmartBundle commercial
software [ 151 ] provides different ways of exploring associations and frequent pat-
terns in market basket data. This software is particularly tailored towards transaction
data. The WizRule software [ 152 ]by WizSoft performs data mining with the use of
association rules. Specifically, this software can also perform classification with the
use of the discovered rules. Thus, this software explores the power of association
rules for data mining in different ways. The XAffinity (TM) software [ 153 ] is suited to
click-stream and Web log data. This software can be used for effective click-stream
and Web log analysis.