This is a WEKA package that wraps the SPMF data mining library (specifically, the association rule and item set mining algorithms).
SPMF is an open-source data mining mining library written in Java, specialized in pattern mining.
It is distributed under the GPL v3 license.
It offers implementations of 88 data mining algorithms for:
* sequential pattern mining,
* association rule mining,
* itemset mining,
* sequential rule mining,
* clustering.
For example, to use SPMF's Apriori algorithm with a min support of 0.4:
java weka.Run weka.associations.SPMFWrapper -M Apriori -P 0.4 -t <file>.arff
The -M parameter specifies the algorithm, and -P the parameters to the algorithm. This link is useful as it lets you know how many parameters each algorithm takes and what those parameters are.
If you have ant, cd into the root directory and run:
ant -Dpackage=spmfWrapper make_package`
Then go to dist and use the WEKA package manager to install spmfWrapper.zip.
- Not all association / item-set mining algorithms work. This is because some of the algorithms take in special datasets that can't be represented in ARFF (or so I believe).
- I wasn't able to "elegantly" pass the currently-loaded ARFF into SPMF so it writes the ARFF to the OS's temp directory and gets SPMF to load it there. That means that on a large dataset things might get quite slow as it has to be saved to disk and then read in again by SPMF. Also, if an exception gets thrown (e.g. you specified the wrong params for an algorithm) the temporary .arff file in the temp folder might not get deleted so this is something I need to try and fix as well.