HGCE Guide
the HGCE algorithm has a number of parameters to it, summarised below:
Usage: gce_serial.py [options] Options: -h, --help show this help message and exit -f FILE, --file=FILE Input filename for edgelist. If not provided, then input will be piped in from STDIN -o OUTFILE, --outfile=OUTFILE Filename for output. If not provided, then found communities will be piped to STDOUT -k MINCLIQUESIZE, --minCliqueSize=MINCLIQUESIZE Minimum size that a clique must have to be considered as a seed. Must be an integer with value 3 or larger. Default: 4. -a ALPHA, --alpha=ALPHA Scaling parameter for GCE (smaller alpha leads to larger comms). Default: 1.0 -e EPSILON, --epsilon=EPSILON Minimum distance that seeds and communities must be from all larger or previously accepted seeds or communities. Default: 0.25 -p PHI, --phi=PHI Minimum proportion of a seed that must not be containted in a previously accepted seed (a.k.a. Clique Coverage Heuristic parameter). Default: 0.25 -i, --intNodeIDs If this flag is included, then node IDs must already be integers (this may save memory---ignore except for huge graphs) -t THREADS, --threads=THREADS Number of threads to run concurrently for clique finding and sorting. Default: 1 -m MINCLIQUEEDGEWEIGHT, --minCliqueEdgeWeight=MINCLIQUEEDGEWEIGHT Minimum weight that an edge must have for it to be eligible for membership in a clique. default = 0.0, -s MAX_TIMES_SPOKEN_FOR, --maxTimesSpokenFor=MAX_TIMES_SPOKEN_FOR Number of times a node/edge must be already included in accepted seeds for it to be considered spoken for. default = 1, -r NUM_REALIZATIONS, --numRealizations=NUM_REALIZATIONS Number of perturbed graphs to create. default = 100, --minAppearanceProp=MIN_APPEARANCE_PROP Proportion of perturbed graphs in which a similar set must appear for that set to be considered significant. default = 0.9, --similarityThreshold=SIMILARITY_THRESHOLD Given two sets, the proportion of items that must be common to both sets for those sets to be considered similar. default = 0.8, -z PERTURB_ALPHA, --perturbAlpha=PERTURB_ALPHA Proportion of edges randomized in perturbed graphs. default = 0.1,