Figure 1
From: Exploiting citation networks for large-scale author name disambiguation
For a given set of publications, a measure for publication similarity is used to identify clusters that ideally represent unique researchers. First, the clustering creates strongly connected components. Second, well-linked clusters are merged.