spacenet.partition.modularity_partition#
- spacenet.partition.modularity_partition(spatial_network, edge_weight_name='Inverse Distance', algorithm='leiden', resolution=1.0, modularity='newman')#
Performs modularity-based community detection on a spatial network using the specified algorithm and parameters. The function computes the modularity partition of the spatial network, which identifies communities of nodes that are more densely connected to each other than to the rest of the network. The edge weights used for computing the modularity partition can be specified through the edge_weight_name parameter, allowing for flexibility in how the community structure is determined based on the spatial relationships between nodes.
- Parameters:
- spatial_networkNetworkX graph
The spatial network for which to compute the modularity partition.
- edge_weight_namestr, optional
The name of the edge attribute to use as the weight for computing the modularity partition. Default is ‘Inverse Distance’.
- algorithmstr, optional
The algorithm to use for community detection. Supported options are ‘leiden’ and ‘louvain’. Default is ‘leiden’.
- resolutionfloat, optional
The resolution parameter for the community detection algorithm. Higher values lead to smaller communities, while lower values lead to larger communities. Default is 1.0.
- modularitystr, optional
The modularity function to optimize. Supported options are ‘newman’, ‘dugue’ and ‘potts’. Default is ‘newman’.
- Returns:
- PartitionResult
A dataclass containing the partitioning results, including:
labels: A dictionary mapping each node to its assigned community.
n_clusters: The number of communities detected in the partition.
modularity: The modularity score of the partition, which is a measure of the strength of the community structure in the graph. Higher modularity values indicate stronger community structure.
Notes
References:
Traag, V. A., Waltman, L., & van Eck, N. J. (2019). From Louvain to Leiden: guaranteeing well-connected communities. Scientific reports, 9(1), 1-12.
Blondel, V. D., Guillaume, J. L., Lambiotte, R., & Lefebvre, E. (2008). Fast unfolding of communities in large networks. Journal of statistical mechanics: theory and experiment, 2008(10), P10008.