Fraud Detection: Ponzi Scheme with Neo4j

Wrote by Marco Falcier, published on his blog.

How to discover a Ponzi Scheme using Neo4j

Banks are wasting billions because of financial frauds. Me, and some other colleagues are working on some bank fraud detection systems using Neo4j.

We first focused on patterns that could easily find some synthetic identities frauds, and next, we moved on discovering new fraud patterns. Our customers asked for a really “popular” fraud case, made by Charles Ponzi, the Ponzi Scheme.

Quoting Wikipedia:

Charles Ponzi

“Charles Ponzi, … was an Italian businessman and con artist in the U.S. and Canada. Born in Italy, he became known in the early 1920s as a swindler in North America for his money-making scheme. He promised clients a 50% profit within 45 days, or 100% profit within 90 days, by buying discounted postal reply coupons in other countries and redeeming them at face value in the United States as a form of arbitrage. In reality, Ponzi was paying early investors using the investments of later investors, a practice known as “robbing Peter to pay Paul.” While this swindle predated Ponzi by several years, it became so identified with him that it now bears his namesake. His scheme ran for over a year before it collapsed, costing his “investors” $20 million.”

Then we were thinking: “What about a graph? Can this schema be converted as a pattern?”.

You can read an analysis with some example here.

Lascia un commento

Il tuo indirizzo email non sarà pubblicato. I campi obbligatori sono contrassegnati *