

INSTALL
	describes the installation procedure

FORMAT
	describes the input and outup format

HOWTO
	Describes how to execute projclos.sh in command line
	and descibes the meaning of the different parameters

data/
	contains sample data sets

data/0README
	describes the 3 sample datasets



------------------------------------------------------



In order to reproduce the results presented in the article submited
to ECML-2015 you could run the following command:

------------------------------------------------------

[-] Musical taste dataset section (ie mougel):
To get the rules obtained for the confuluence wuth  ”support≥ 4“ in
the set of friends O you should run

./proj_clos.sh      ${DIR_PROJCLOS}/data/mougel.ri  \
                    Local 4 0 0 4                   \
                    ${DIR_PROJCLOS}/data/mougel.abs

------------------------------------------------------

[-] Teenage Friends and Lifestyle Study’s dataset (ie s50)

./proj_clos.sh      ${DIR_PROJCLOS}/data/s50.ri  \
                    Local 4 0 0 4                   \
                    ${DIR_PROJCLOS}/data/s50.abs

------------------------------------------------------
[-] In order to test the programme on a larger dataset you may
want to run it on a dataset dblp.ri (32 items in the language
of description of the itemsets, 45131 vertices, 228188 edges)
with a frequent global closed pattern minimal support of 1% (ie 
452), "min_degree >= 16"Abstraction and "support>=4"confluence
in the graph of 3-cliques. Computing frequent gloabl pattern with
CORON is very fast but computing local abstract closed pattern
can take few minutes.

./proj_clos.sh      ${DIR_PROJCLOS}/data/dblp.ri  \
                    Local 452 0 0 16              \
                    ${DIR_PROJCLOS}/data/dblp.abs



