GcGrowth, Version 1.
====================

GcGrowth is a fast algorithm for mining equivalence classes of
frequent patterns. The mined equivalence classes are represented
as convex spaces bounded by generators and closed patterns.


Usage:
	GcGrowth data_filename min_sup(an absolute number) [output_filename]


data_filename: Each line represents an object with a set of attributes 
               (integers) seperated by space, or a transaction containing
               a set of items. See sample input file "mushroom.dat".  


min_sup: the minimum support threshold, it is an integer. 


Output_filename: three files are generated. If the output_filename is 
                 "pat", then the three files are "pat.closed", "pat.key"
                 and "pat.pair". File "pat.key" contains the generators. 
                 Its format is as follows: every line represents a generator.
                 The first integer is the length of the generator, i.e.,
                 the number of items contained in the generator. The 
                 following integers are the set of items contained in 
                 the generator. The last integer is the support of the itemset.
                 The format of the "pat.closed" is the same as that of "pat.key".
                 File "pat.pair" contains information that links generators and 
                 closed patterns in the same equivalent class together. Each line
                 contain a pair of ids. The first id is a generator id and 
                 the second id is the id of the closed pattern that is in the same
                 equivalent class as the generator. For generators and closed
                 patterns, their ids are their line no in the corresponding file. 

Example: GcGrowth mushroom.txt 81 pat

Credits:

This program was developed by Guimei Liu. This project is
supported in part by the I2R-SOC Joint Lab on Knowledge Discovery
and FRC grants "R-252-040-238-101 & R-252-060-238-133: 
Pattern Spaces: Theory, Algorithms, and Applications".

If you use these programs, please cite the following papers:

Haiquan Li, Jinyan Li, Limsoon Wong, Mengling Feng, Yap-Peng Tan. 
Relative Risk and Odds Ratio: A Data Mining Perspective. 
Proceedings of 24th ACM SIGMOD-SIGACT-SIGART Symposium on 
Principles of Database Systems, 
pages 368--377, Baltimore, Maryland, June 2005.
http://www.comp.nus.edu.sg/~wongls/psZ/sub-217-li.pdf