missSBM: Handling missing data in Stochastic Block Models

Travis build status AppVeyor build status Coverage status

When a network is partially observed (here, missing dyads, that is, entries with NA in the adjacency matrix rather than 1 or 0), it is possible to account for the underlying process that generates those NAs. missSBM is an R package for adjusting the popular Stochastic Block Models from network data sampled under various missing data conditions.




Please cite our work using the following reference:

Timothée Tabouy, Pierre Barbillon & Julien Chiquet (2019) “Variational Inference for Stochastic Block Models from Sampled Data”, Journal of the American Statistical Association, DOI: 10.1080/01621459.2018.1562934