SNetOC

Matlab package for Sparse and modular Networks with Overlapping Communities (SNetOC)

Version: v1.0
Last modified: 2017-11-16
URL: https://github.com/misxenia/SNetOC

Description

This Matlab package implements algorithms for simulation and posterior inference with the class of sparse graph models with overlapping community structure introduced by Todeschini, Miscouridou and Caron (2017). It allows to simulate graphs with a given level of sparsity and a fixed number of communities. It also performs inference of the network parameters (level of affiliation of each node to each community) and hyperparameters (sparsity of the network).

The package has been tested on Matlab R2016a, R2017a and R2017b and requires the Statistics Toolbox.

Reference

A. Todeschini, X. Miscouridou, F. Caron, Exchangeable Random Measures for Sparse and Modular Graphs with Overlapping Communities. arXiv:1602.02114. Download paper.

Installation

Download the zip file
Unzip it in some folder
Run the test file test.m

In order to use the package, the folders 'GGP', 'CGGP' and 'utils' need to be added to the Matlab path, using the command addpath (see the test file).

Test script

test.m: checks that the basic functions are running without error

Demos and reproducible results of the article Todeschini, Miscouridou and Caron (2017)

demo_sparsity.m: shows empirically the sparsity properties of a range of graph models. [html version]

You can run multiple MCMC chains in parallel by calling the parpool command before running the following scripts. This requires the Parallel Computing Toolbox.

demo_simulations.m: posterior inference on a simulated graph under the CCRM model. [html version]
demo_polblogs.m: posterior inference on the polblogs graph under the CCRM model. [html version]
demo_usairport.m: posterior inference on the USairport graph under the CCRM model. [html version]
demo_overlappingcommunity.m: posterior inference on the polblogs graph under the CCRM model using the wrapper function overlapping_community_detection. [html version]
MMSB_polblogs.m: posterior inference on the polblogs graph the mixed membership stochastic blockmodel.
MMSB_usairport.m: posterior inference on the USairport graph the mixed membership stochastic blockmodel.
MMSB_simulations.m: posterior inference on a simulated graph under the mixed membership stochastic blockmodel.

Main functions: simulation and posterior inference on graphs

overlapping_community_detection: Wrapper function; takes a graph as input and returns the level of affiliation of each node to different communities
graphmodel: creates a graph model object
graphrnd: samples a graph
graphmcmc: creates a MCMC object
graphmcmcsamples: runs a MCMC algorithm for posterior inference on graphs
graphest: returns point estimates of the graph parameters

Changes

v1.0 (2017-11-16)

First release of the package.

Copyright

SNetOC is Copyright (c) 2016-2017 A. Todeschini (adrien.todeschini@gmail.com), X. Miscouridou (xenia.miscouridou@spc.ox.ac.uk) and F. Caron (caron@stats.ox.ac.uk). All Rights Reserved. See the file LICENSE.txt.

blizzardwj / SNetOC