This Apriori-based implementation is a part of the FIM template library. If I change the code, then I change the version found in the FIM template Library. Therefore, the version that can be downloaded from this page, is not updated any more and this page is no longer maintained. Please click here to go to the webpage of the FIM template library.
Apriori algorithm was originally proposed by Agrawal in "Fast Algorithms for Mining Association Rules" in 1994 to find frequent itemsets (known as the problem of FIM) and association rules in a transaction database. A natural generalization of FIM is finding frequent sequence of items in a transactional database where both the transactions and the patterns are allowed to contain duplicates. To solve this problem the trie-based Apriori algorithm can easily be extended.
Here you can download a fast, trie-based, command-line implementation of the Apriori algorithm for Linux platform.
The code is written in standard C++ in object-oriented manner.
It uses STL possibilities, if that does not reduce running speed.
The code can be freely used for research purposes.
If you use this program in a research paper then a citation to the following paper is welcome.
To compile in Unix systems (obviously after uncompressing the downloaded file) do the following:
cd src; make dep; make ../fsm; cd ..
this will result in an fsm program. To generate the documentation type
cd src; make doc; cd ..
To run fsm 4 parameters are mandatory:
For example you can type in a linux terminal:
./fsm apriori kosarak2_10_2.dat 2 out.txt
or under Windows command prompt:
.\fsm apriori kosarak2_10_2.dat 2 out.txt