cdlib.algorithms.agdl

agdl(g_original, number_communities, number_neighbors, kc, a)

AGDL is a graph-based agglomerative algorithm, for clustering high-dimensional data. The algorithm uses the indegree and outdegree to characterize the affinity between two clusters.

Parameters:
  • g_original – a networkx/igraph object
  • number_communities – number of communities
  • number_neighbors – Number of neighbors to use for KNN
  • kc – size of the neighbor set for each cluster
  • a – range(-infinity;+infinty). From the authors: a=np.arange(-2,2.1,0.5)
Returns:

NodeClustering object

Example: 

>>> from cdlib import algorithms
>>> import networkx as nx
>>> G = nx.karate_club_graph()
>>> com = algorithms.agdl(g, number_communities=3, number_neighbors=3, kc=4, a=1)
References:

Zhang, W., Wang, X., Zhao, D., & Tang, X. (2012, October). Graph degree linkage: Agglomerative clustering on a directed graph. In European Conference on Computer Vision (pp. 428-441). Springer, Berlin, Heidelberg.

Note

Reference implementation: https://github.com/myungjoon/GDL