Thanks to visit codestin.com
Credit goes to github.com

Skip to content

MrDragar/xgi

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

XGI

What has been done?:

Metrics computation has been updated

The computation of the linegraph vector centrality metric was accelerated by using the igraph library instead of networkx . For comparison, implementations using the networkx.Graph class were retained.

H_enron = xgi.load_xgi_data("plant-pollinator-mpl-034")
H_enron_cleaned = H_enron.cleanup(
    multiedges=False, singletons=False, isolates=False, relabel=True, in_place=False
)
start_time = time.time()

xgi.centrality.line_vector_centrality(H_enron_cleaned)

middle_time = time.time()

xgi.centrality.fast_line_vector_centrality(H_enron_cleaned)

end_time = time.time()

print(f"Время выполнения nx: {(middle_time-start_time):.6f} секунд")
print(f"Время выполнения ig: {(end_time-middle_time):.6f} секунд")

The implementation of bipartite graphs has been changed.

The implementation for working with bipartite graphs was modified: the igraph.Graph class is now used instead of networkx.Graph for converting from a Hypergraph to a bipartite graph and back to a Hypergraph .

hg = xgi.Hypergraph()
hg.add_edges_from(
    [[1, 2, 3], [3, 4, 5,8], [1, 4, 10, 11, 12], [7,0,8], [5,7,2]]
)

bpg = xgi.convert.to_bipartite_graph(hg)
plot(bpg,vertex_label=bpg.vs.indices,layout = Graph.layout_bipartite(bpg), bbox=(0, 0, 500,500))

Application of graph and hypergraph metrics to improve ML models

In the file for solving the classification problem, in addition to article embeddings, article identifiers, and author IDs used as features, metrics such as eigenvector centrality of the graph and vector centrality of the hypergraph were also utilized.

Cpp implementation of linegraph

  • In progress...

About

CompleX Group Interactions (XGI) is a Python package for higher-order networks.

Resources

License

Code of conduct

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Jupyter Notebook 90.9%
  • Python 9.1%