communityalg

Algorithms and functions in Matlab for community detection in networks. Expands BrainConnectivity toolbox.

Matlab/Octave algorithms and functions

• ami.m Returns the adjusted mutual information between two membership vectors.
• association_score.m Returns the association score between pairs of communities specified by the graph and membership.
• asymptotic_modularity.m Compute asymptotic modularity of a graph with respect to a membership vector.
• asymptotic_modularity_sum.m TODO
• asymptotic_surprise.m Compute asymptotic surprise of a graph with respect to a membership vector.
• clique.m Generate an adjacency matrix of a clique graph with n nodes.
• cluster_similarity.m Compare two membership vectors.
• clustering_entropy.m Compute the clustering entropy of an agreement matrix as in "Gfeller, Newman, 2006".
• community_robustness_weighted.m TODO
• community_size2memb.m Convert an array where every element is the size of a clique to the correspondig membership vector.
• comm_mat.m Returns the block matrix of a graph and its community structure as membership.
• compute_surprise.m Compute the surprise given surprise paramenters.
• consensus_clustering.m TODO
• consensus_clustering_weighted.m TODO
• consensus_entropy.m TODO
• consensus_robustness.m TODO
• correlation_louvain.m Adaptation of the BCT community_louvain method for correlation matrices, as described in MacMahon,2015.
• count_comm.m Plot the histogram with the community size given a membership vector.
• cycle_graph.m Generate the adjacency matrix of a cycle graph with n nodes.
• effcommplot.m TODO
• generate_agreement.m Generate the agreement matrix for a given community detection method.
• generate_agreement_weighted.m Generate the weighted agreement matrix for a given community detection method.
• generate_connected_components.m TODO
• graph_JS_similarity.m Compute the quantum Jensen Shannon divergence between two adjacency matrices.
• graph_laplacian.m Compute the graph combinatorial Laplacian matrix L=D-A.
• group2membership.m Convert a cell of arrays representing the nodes in the communities to a membership vector.
• image_to_network.m Convert a gray index image to its corresponding adjacency graph.
• imagesctxt.m Show a matrix like imagesc but with text values of elements displayed on the pixels.
• isoctave.m Returns true if using Octave, false if using Matlab.
• jensen_shannon_sim.m Returns the Jensen-Shannon symmetrized information theoretic distance between two graph Laplacians.
• k_regular.m Generate a k-regular graph, a graph where the degree of every vertex is k.
• KL.m Returns the binary Kullback-Leibler divergence between Bernoulli distribution p and q.
• kullback_leibler_sim.m Returns the Kullback-Leibler divergence between two graph Laplacians.
• logHyperProbability.m Compute the logarithm of the hypergeometric probability in base 10.
• membership2groups.m Convert a membership vector to a cell of arrays of nodes in every community.
• membership_agreement.m DEPRECATE
• membership_similarity.m TODO
• method_best.m Functio handle to the community detection method that returns the best value over a set of repetitions.
• modularity.m Returns the modularity of a graph with respect to a membership vector.
• nearcorr.m Returns the nearest correlation matrix of a square matrix. Implementation by Nick Higham.
• nearestSPD.m Returns the positive definite matrix of a square matrix. Implementation by Nick Higham.
• norm_conf_mat.m DEPRECATE
• number_connected_components.m Returns the number of connected components of a graph.
• number_of_edges.m Returns the number of edges of a binary or weighted graph.
• paco.m Function handle to the MEX implementation of PACO.
• partition_params.m Returns the partition parameters for use with compute_surprise.
• quantum_density.m Returns the quantum density of a graph, Brauenstein et al. "Ann. of Combinatorics, 10, no 3 (2006), 291-317."
• reindex_membership.m Transform a membership vector to have community indices sorted by community size from 1 to |C|
• reorder_membership.m Linearize a membership vector to have continuous indices of communities from 1 to |C|
• ring_of_cliques.m Returns a network ring of cliques, with given number of cliques and clique size and its membership.
• ring_of_custom_cliques.m Returns a network ring of cliques, with given size of cliques specified as input and its membership.
• rmtdecompose.m Returns the Random Matrix Theory decomposition of a correlation matrix.
• robustness_configuration_interp_und.m
• robustness_edge_weight_und.m
• run_cluster_similarity.m
• rwalkent.m Returns the random walk entropy of a graph as in Estrada et al. Walk entropies in graphs, "Linear Algebra and its Applications 443 (2014) 235–244"
• significance.m Returns the significance of a graph partitioning, Traag (2013).
• smi.m Returns the standardized mutual information of two membership vectors.
• sort_group_by_size.m Sort community groups by size.
• star_of_custom_cliques.m Returns a star of cliques, every clique is connected to all other cliques with one edge.
• surprise.m Returns the surprise of a graph partitioning.
• threshold_by_giant_component.m Returns the threshold over which the graph has more than one connected component.
• threshold_by_num_edges.m Returns a graph thresholded to have a specific number of edges.
• vonneumann_entropy.m Returns the VonNeumann quantum entropy of graph Brauenstein et al. "Ann. of Combinatorics, 10, no 3 (2006), 291-317."
• write_brainet.m Write a graph with coordinates of nodes and membership to Brainet format.
• write_brainet_community.m TODO
• writetoEdgesList.m TODO
• writetoPAJ_labels.m
• writetoPAJ_labels_coords.m

Python algorithms

• find_intersections.py Find the intersections of edges in a bipartite graph (TODO)

Datasets

This repository contains the 638 areas template used by Crossley (2013). It consists of two files:

1. template.nii is the NIFTI file describing the template, MNI space
2. template_638.txt is the description of anatomical a

Additionally the file template_638_coords_abbr.txt is organized as follows

Node	X	Y	Z	Label	Abbr

and differently from template_638.txt the NodeID starts from 0 to 637 (for indexing with Python).