Chinese Whispers for Python¶

This is an implementation of the Chinese Whispers graph clustering algorithm in Python.

  • https://pypi.org/project/chinese-whispers/
  • https://github.com/nlpub/chinese-whispers-python

Version Information¶

In [1]:
from chinese_whispers import __version__ as cw_version
from networkx import __version__ as nx_version
from matplotlib import __version__ as plt_version
print('Chinese Whispers {}'.format(cw_version))
print('NetworkX {}'.format(nx_version))
print('matplotlib {}'.format(plt_version))

Chinese Whispers 0.8.2.rc1
NetworkX 3.1
matplotlib 3.7.1

Clustering¶

In [2]:
import networkx as nx
from chinese_whispers import chinese_whispers, aggregate_clusters
In [3]:
G = nx.karate_club_graph()
In [4]:
# Perform clustering of G, parameters weighting and seed can be omitted
chinese_whispers(G, weighting='top', seed=1337) 

# Print the clusters in the descending order of size
print('ID\tCluster\n')

for label, cluster in sorted(aggregate_clusters(G).items(), key=lambda e: len(e[1]), reverse=True):
    print('{}\t{}\n'.format(label, cluster))

ID	Cluster

24	{32, 33, 8, 9, 14, 15, 18, 20, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31}

3	{0, 1, 2, 3, 4, 7, 10, 11, 12, 13, 17, 19, 21}

7	{16, 5, 6}

Visualization¶

In [5]:
%config InlineBackend.figure_formats = ['retina']

import matplotlib.pyplot as plt
In [6]:
colors = [1. / G.nodes[node]['label'] for node in G.nodes()]

nx.draw_networkx(G, cmap=plt.get_cmap('jet'), node_color=colors, font_color='white')

Citation¶

  • Ustalov, D., Panchenko, A., Biemann, C., Ponzetto, S.P.: Watset: Local-Global Graph Clustering with Applications in Sense and Frame Induction. Computational Linguistics 45(3), 423–479 (2019)
@article{Ustalov:19:cl,
  author    = {Ustalov, Dmitry and Panchenko, Alexander and Biemann, Chris and Ponzetto, Simone Paolo},
  title     = {{Watset: Local-Global Graph Clustering with Applications in Sense and Frame Induction}},
  journal   = {Computational Linguistics},
  year      = {2019},
  volume    = {45},
  number    = {3},
  pages     = {423--479},
  doi       = {10.1162/COLI_a_00354},
  publisher = {MIT Press},
  issn      = {0891-2017},
  language  = {english},
}

In case you require higher performance, please consider our Java implementation that also includes other graph clustering algorithms: https://github.com/nlpub/watset-java.

In [ ]: