Community Plus

class rdigraphs.community_plus.community_plus.CommunityPlus

Bases: object

Generic class for the detection and analysis of communities from graphs

community_metric(edges, weights, clabels, parameter)

Computes a community metric with respect to a given graph

Parameters:

edges (list of tuples of int) – List of edges, as pairs of indices (i, j)
weights (list of float) – Weight of each edge
clabels (list) – List of community labels
cd_alg (str) – Name of the community detection algorithm
parameter (str) – Name of the global parameter to compute. Available options are: ‘coverage’, ‘performance’ and ‘modularity’

Returns:

q – Value of the selecte metric

Return type:

float

compare_communities(comm1, comm2, method='vi', remove_none=False)

This is a wrapper to methods compare_communities and split_joing_distance from igraph.

Compares two community structures using various distance measures.

Parameters:

comm1 (1st community structure as a membership list)
comm2 (2nd community structure as a membership list)
method (str) – Measure to use. Options are:

‘vi’ | ‘meila’ : variation of information metric [Meila]

‘nmi’ | ‘danon’ : normalized mutual information [Danon]

‘rand’ : Rand index [Rand],

‘adjusted_rand’ : means the adjusted Rand index [Hubert]

‘split-join’ : split-join distance [van Dongen]

‘split-join-proj’ : assymetric split-join distance [van Dongen]
remove_none (boolean) – Whether to remove None entries from the membership lists.

Return type:

The calculated measure.

References

Meila M: Comparing clusterings by the variation of information. In:: Scholkopf B, Warmuth MK (eds). Learning Theory and Kernel Machines: 16th Annual Conference on Computational Learning Theory and 7th Kernel Workship, COLT/Kernel 2003, Washington, DC, USA. LCNS, vol. 2777, Springer, 2003. ISBN: 978-3-540-40720-1.
Danon L, Diaz-Guilera A, Duch J, Arenas A: Comparing community: structure identification. J Stat Mech P09008, 2005.
van Dongen D: Performance criteria for graph clustering and Markov: cluster experiments. Technical Report INS-R0012, National Research Institute for Mathematics and Computer Science in the Netherlands, Amsterdam, May 2000.
Rand WM: Objective criteria for the evaluation of clustering methods.: J Am Stat Assoc 66(336):846-850, 1971.
Hubert L and Arabie P: Comparing partitions. Journal of Classification: 2:193-218, 1985.

Notes

‘vi’ and ‘split-join’ are divergence measures (minimum at 0). ‘nmi’, ‘rand’ and ‘adjusted_rand’ are similarity measures (max at 1).

detect_communities(edges, weights, n_nodes=None, alg='louvain', ncmax=None, resolution=1e+100)

Applies a Community Detection algorithm to the graph given by a list of edges and a list of weights.

Parameters:

edges (list of tuples of int) – List of edges, as pairs of indices (i, j)
weights (list of float) – Weight of each edge
n_nodes (int) – Total number of nodes. In general, it must be specified because isolated nodes do not appear in the list of edges. If none, it is taken as the largest index in list or edges
ncmax (int) – Number of clusters.
alg (string, default=’louvain) – Community detection algorithm. It must be one of ‘louvain’, ‘spectral’, ‘fastgreedy’, ‘walktrap’, ‘infomap’ or ‘labelprop’
resolution (float) – Resolution parameter of the Louvain algorithm

Returns: