Labelpropagation: Graph Community Detection for Large-Scale Networks

Community labels for large-scale networks are generated by this 2022 implementation of the Raghavan et al. algorithm by Benedek Rozemberczki. The resource provides a near-linear time framework for unsupervised graph clustering within the NetworkX ecosystem.

Use Cases

• Community detection in social graphs using the label propagation function to identify node clusters
• Analyzing homophily in networks by partitioning nodes into unsupervised groups
• Comparing modularity outcomes against Louvain or Walktrap clustering methods

Strengths

• Implements a peer-reviewed algorithm from Physical Review E 2008
• GPL-3.0 licensed for open-source use
• Compatible with NetworkX graph objects

Limitations

• No pre-packaged graph data included in the metadata
• Randomized nature of the algorithm can result in different community structures per execution
• Temporal staleness as the repository has not been updated since late 2022

Provenance

Source

Physical Review E (2008) publication by Raghavan et al.

Collection Method

Algorithm implementation

Freshness

Last updated November 2022; no ongoing update frequency specified.