Crypto / Digital-Asset Securities methodology →

EDGAR forms: 1-A · S-1 · 8-K

recall 43% precision 2%

Financial-Statement Fraud methodology →

EDGAR forms: 10-K · 10-Q · 8-K · NT 10-K

recall 67% precision 23% + content rule precision 26%

Insider Trading methodology →

EDGAR forms: 4 · 3 · 5 · 144 · 8-K · 10-Q

recall 81% precision 19%

Broker-Dealer Fraud methodology →

EDGAR forms: Form D

recall 29% precision 9%

Boiler Room methodology →

EDGAR forms: Form D

recall 50% precision 4%

FCPA Violation methodology →

No EDGAR filing fingerprint — criminal/DOJ-side scheme, detected from case text.

Securities Anomalies

Full Corpus Statistics 3,648

SEC/DOJ Heuristics

3,648 documents, 3 sources, 944.6MB, 1995–2026. Filtered to securities-fraud signals (pump-dump · unregistered · ticker · manipulation · reverse-merger · boiler-room · shell · microcap · penny-stock · insider).

Selected scheme types: insider-trading, financial-fraud, broker-dealer-fraud, boiler-room, crypto-securities, fcpa contains 3,648 documents clear

SEC 3,402 docs

SEC Litigation Releases / complaint 659 | 1995–2026 | 534.0MB

SEC Litigation Releases / judgment 52 | 2011–2026 | 28.0MB

SEC Litigation Releases / litigation release 1,918 | 1995–2026 | 122.7MB

SEC Press Releases / complaint 215 | 2006–2026 | 201.7MB

SEC Press Releases / press release 558 | 1997–2026 | 26.2MB

DOJ 246 docs

DOJ SDNY / indictment 5 | 2013–2024 | 3.7MB

DOJ SDNY / press release 241 | 2013–2026 | 28.3MB