Machine Learning for Money Laundering Analysis
Six supervised models on the Elliptic Bitcoin dataset were compared; Random Forest achieved 98.87% accuracy on the holdout set.
AUSTRAC
Regulator · AU · 2 documents
Key number
98.87%
Random Forest accuracy on the holdout set. International Journal of Network Dynamics and Intelligence p.22
Key number
5-10%
Average annual turnover consumed by regulatory divergence. Revista Gênero e Interdisciplinaridade p.5
International perspectives on de-risking
Regulatory divergence creates material and growing expenses in the global financial sector, taking up on average 5-10% of annual turnover.