tnetwork.DCD.label_smoothing¶

tnetwork.DCD.label_smoothing(dynNetSN, CDalgo='louvain', match_function=<function jaccard>, threshold=0.3, multithread=False, **kwargs)[source]¶

Community detection by label smoothing

This method is based on falkowsky et al.[1]. It first detect communities in each snapshot, then try to match any community with any other one in any other snapshot, constituting a survival graph. A community detection algorithm is then applied on this survival graph, yielding dynamic snapshot_communities.

[1]Falkowski, T., Bartelheimer, J., & Spiliopoulou, M. (2006, December). Mining and visualizing the evolution of subgroups in social networks. In Proceedings of the 2006 IEEE/WIC/ACM International Conference on Web Intelligence (pp. 52-58). IEEE Computer Society.

Parameters:

dynNetSN – a dynamic network
CDalgo – community detection to apply at each step. Can be a function returning a clustering, or the string “louvain” or “smoothedLouvain”
match_function – a function that gives a matching score between two snapshot_communities (two sets of nodes). Default: jaccard
threshold – a threshold for match_function below which snapshot_communities are not matched

Returns:

DynCommunitiesSN