Further, the similarity between the added and removed bicliques can be computed using the Jaccard distance to facilitate a better understanding of the structural changes and the influence of missing links. In the Motifs Detail View 302, when an analyst hovers over a biclique, this information is shown as links connecting the related bicliques, with the thickness of the links mapped to their pairwise similarity value.

Computing node-metrics is a method used for getting a picture of the characteristics of a network in social sciences and other domains. The Metrics View 304 in interface pane (e) supports this kind of analysis by presenting a number of metrics in a traditional tabular view, including the degree, closeness, and betweenness centralities of before and after adding certain missing links. Changes of metric values are highlighted (e.g., in red). This table is also interactively linked with other views. For example, hovering over a row emphasizes the corresponding node in the Network View 301. As there might a large number of nodes (rows), a search function can also be provided, and hovering over a node in other views automatically navigates to the corresponding row in the table.

To validate the accuracy of the proposed missing link prediction approach, quantitative experiments were conducted with three bipartite networks, including a weighted person-place network extracted from the Atlantic Storm corpus, a weighted user-conversation bipartite network detected from Slack communication messages, and an unweighted bipartite network between authors and papers from the IEEE VIS publication corpus.