Branch: master
Find file History
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
..
Failed to load latest commit information.
LGL
Makefile
PROXY
README
agents
analysis
centr.m
chkcorpus
clusco.m
clusco_single.m
connect.m
debug.m
decomp.m
eval_basic.m
eval_exp.m
eval_maxit.m
eval_threshold.m
get
getagent
getproxy
load_data.m
load_diag.m
matlab
meta.slashdot-zoo
mkconfig
mkcorpus
mkedges
mkeval
mkgraph
mklayout
mklayoutpng
mkmatrix
mkmatrixclean
mkname
mknet.cc
mkpred
mksplit
mkuserids
near.m
pca.m
pca_draw.m
pca_draw_one.m
plfit.m
pred.m
pred_full.m
pred_sparse.m
progress
progress.eps
progress2.eps
proxies
resolve-userids
save_diag
scatter.m
set_sparam.m
sleep
sort_ud.m
spectral_transformation.m
stat.m
track-progress
troll.m
troll_eval.m
wget_args

README

This is the extraction and analysis code for the Slashdot Zoo. 

== Crawling ==

The zoo is extracted to $ZOO, which defaults to ~/zoodir.  This
zoodir contains the extracted data with one dir per user ($ZOO/user.*),
containing four files (friends, fans, foes, freaks).  Usernames are URL
encoded. 

Crawling is done by executing "./get".  This is a script which crawls
continuously and which can be interupted and restarted when necessary. 

Requirements:  perl, wget 

=== Important files ===

get -- the extraction proper
analysis -- statistics
progress -- progress chart
mk* -- scripts to generate the corresponding out.* file


== Analysis ==

Analysis of the Slashdot Zoo is done using a Matlab, and the Large
Graph Library LGL (version 1.1 works). 

Files can be generated using the Makefile. 

out.* -- files generated automatically
out.corpus -- The whole corpus with one relation per line
out.userids -- mapping between usernames and userids.  Userids begin at
	1 and are continuous.
out.matrix -- The whole corpus in Matlab readable format (one relation
	per line, using userids and +1/-1 as relation).
out.stat -- Statistics generated using Matlab
out.pred -- Prediction statistics
*.{ps,eps,png} -- figures generated automatically

=== Requirements ===

- GNU Make
- Matlab:  You need "matlab" in the path
- LGL-1.1:  Large Graph Layout, see file LGL

=== Environment ===

$verbose=1	output more
$fast=1		be faster but less accurate