SPNSA (Shortest Paths Network Search Algorithm)

A feed-guided subgraph-extraction method: seed it with a small set of "known bad" entities (the feed) and it searches outward along shortest paths to surface a compact, interpretable sub-network connecting them. Introduced in 2015 to find Enron's convicted money-launderers inside the Enron e-mail corpus, it was carried eleven years later into financial anti-money-laundering (AML) research — applied to the Elliptic Bitcoin graph and the synthetic IBM-AML benchmark family, with a new 2026 paper proposing better strategies for choosing the feed itself rather than changing the algorithm. The vault's interest in SPNSA is specifically that cross-domain migration: a single research lineage, carried by a shared co-author, moving an investigative tool from one crime domain's dataset to an entirely different one's.

References

• claim-spnsa-originated-2015-on-enron-email-corpus-not-financial-transactions
• claim-spnsa-motif-feed-selection-outperforms-baselines-but-not-universally
• claim-spnsa-illicit-enrichment-gain-traces-to-feed-selection-not-path-aggregation