graphpca 0.3

Produces a low-dimensional representation of the input graph

Produces a low-dimensional representation of the input graph.

Calculates the ETCD [1] of the graph and reduces its dimension using PCA. The result is an embedding of the graph nodes as vectors in a low-dimensional space.

Usage

>>> import networkx as nx
>>> import graphpca
>>> g = nx.erdos_renyi_graph(10000, 0.2)
>>> g_2 = graphpca.reduce_graph(g, 2)
>>> graphca.plot_2d(g_2)

Contributing

Feel free to fork me and create a pull request at https://github.com/brandones/graphpca

[1]

https://www.info.ucl.ac.be/~pdupont/pdupont/pdf/ecml04.pdf