Automatic Dimension Selection for a Non-negative Factorization Approach to Clustering Multiple Random Graphs

This work addresses clustering of multiple graphs, which is useful for network analysis, but appears incremental as it builds on existing factorization methods.

The authors tackled the problem of grouping multiple graphs into clusters by developing a model selection information criterion to automatically estimate the number of clusters, using singular value thresholding and non-negative factorization, and demonstrated performance comparable to standard algorithms on simulated and Swimmer datasets.