Harmonic Centrality vs. Closeness Centrality

Classic closeness centrality is commonly defined using the inverse of a node’s average distance to other nodes. That is straightforward when every node can reach every other node, but disconnected graphs require special handling.

Harmonic centrality instead sums reciprocal distances. Unreachable nodes simply contribute zero. It therefore gives a natural score in directed or disconnected networks without inventing a distance for nodes that cannot be reached.

Both are distance-based centrality measures, but their formulas and behavior differ. Continue with the harmonic-centrality formula or see how it differs from PageRank.