msakai / plcm Goto Github PK

PLCM is a frequent itemset mining algorithm based on the LCM algorithm. 

##################
#### Download ####
##################

Latest version at: http://people.cs.kuleuven.be/~benjamin.negrevergne/plcm/plcm_latest.tar.gz

##################################
##### Reference publications #####
##################################

Discovering Closed Frequent Itemsets on Multicore:
Parallelizing Computations and Optimizing Memory Accesses. 

Benjamin Negrevergne Alexandre Termier Jean-François Méhaut And
Takeaki Uno.

LCM ver. 2: Efficient Mining Algorithms for
Frequent/Closed/Maximal Itemsets

Takeaki Uno1, Masashi Kiyomi, Hiroki Arimura2

###################
##### Compile #####
###################

./configure 
make
make install 

### debug mode ###

./configure --enable-debug=yes
make
make install 

#################
###### Run ######
#################

plcm <dataset file> <absolute threshold> <output file> [ -t number of threads ]

(If output is '-' then plcm dumps itemsets on the standard output.)

#######################
##### File format #####
#######################

### Input file ###
Dataset must be a single ASCII file. 

Each line is a transaction. 
Each transaction contains distinct items. 
Transaction must be ordered. 
Last line must be empty.

eg. 
--- file test.dat ---
1 2 4 6
1 2 3 5 7
2 3

is a valid dataset. 

### Output file ###

plcm output files have the following format: 
* each line is a frequent closed itemset
* frequency is stored at the end of the line into brackets.

eg.
running plcm on test.dat dataset, with the following command 
./plcm test.dat 2 out.dat -t 1

will generate the following file 

--- file out.dat ---
2 (3)
1 2 (2)
3 2 (2)

You can change the output format by modifying the dumpItemset()
function in plcm.hpp and plcm.cpp

################
##### BUGs #####
################

Report bugs and/or comments at: 
[email protected]

My FirstName is Benjamin
My LastName is Negrevergne

If you observe a bug, please compile in debug mode and report the error message if any. 
(To compile in debug mode use ./configure --enable-debug=yes.)


######################
##### Developers #####
######################

To checkout the source tree:
git clone <plcm_git_repos> plcm
cd plcm
git submodule init
git submodule update

Compile with the standard procedure (see Compile section)

./configure 
make 
make install