From: Inferring Missing Links from the Paths That Remain

If you had a partially observed network, which missing connections would you investigate first?

The question forces you to separate high structural plausibility from high practical value. A predicted edge may be statistically interesting but irrelevant, while a modestly ranked connection may matter greatly in a safety, research, or organisational setting.

Action

Choose a small network, score absent pairs by common neighbours, and manually inspect the top five.

Read the full exploration

What else is in this exploration

3 evidence blocks3 perspectives2 visualizations8 media resources4 rabbit holes

evidence

Link prediction estimates which unobserved pairs of nodes are likely to be connected.

evidence

Hierarchical structure can improve prediction of missing network connections.

evidence

Common-neighbour counts provide a basic path-based score for a possible missing link.

Continue exploring

Inferring Missing Links from the Paths That Remain

Evidence, perspectives, rabbit holes, and more