ostroumova-la/cascade_communities

Cascade-based community detection

This is a supplementary code for the paper When Less is More: Systematic Analysis of Cascade-based Community Detection.

Supplementary directories:

CommDiff-Package/: the source code for R-CoDi and D-CoDi from the paper “Community Detection Using Diffusion Information” by Ramezani et al.

CommunityWithoutNetworkLight/: the source code for C-IC and C-Rate algorithms from the paper “Efficient methods for influence-based network-oblivious community detection” by Barbieri et al.

network-inference-multitree/: the source code for the MultiTree algorithm from the paper “Submodular inference of diffusion networks from multiple trees” by Gomez-Rodriguez et al.

community_ext/: community detection library complementing the paper “Community detection through likelihood optimization: in search of a sound model” by Prokhorenkova et al.

benchmark/: used to generate synthetic graphs according to the LFR model proposed in “Benchmark graphs for testing community detection algorithms” by Lancichinetti et al.

Datasets directories:

LFR_1000/, citeseer/, cora-small/, cora/, dolphins/, eu-core/, football/, karate/, newsgroup/, polblogs/, polbooks/, twitter/.

Directories with results: average_ranks/ and average_results/ contain aggregated results over real-world datasets, cascade_plots/ contains the distribution of cascade sizes.

Description of scripts:

C-SI-BD.py, SI-BD.py, SIR.py - to generate epidemics;

base_algorithms.py, base_algorithms_twitter.py - simple algorithms;

baseline_barbieri.py, baseline_barbieri_twitter.py - to run C-IC and C-Rate;

baseline_cd.py, baseline_cd_twitter.py - to run R-CoDi and D-CoDi;

opt_algorithms.py, opt_algorithms_twitter.py - GraphOpt and ClustOpt algorithms;

cascade_plots.py, cascade_plots_twitter.py - to generate data for plots;

average_rank.py, average_results.py to aggregate the results.

To reproduce the main experiments from the paper one can use the file paper_experiments.tex.