Friday, February 17, 2017

Shinhan Bank, introduced deep learning-based 'detection systems over financial transactions'



[Korea sinyuncheol financial newspaper reporter] said 17 days Shinhan Bank has introduced 'financial transactions over detection system, first apply deep learning financial skills.

Shinhan Bank is the last in 2013 by analyzing trading patterns fraud type deal together introduced the "over financial transactions detection system (Fraud Detective System) 'for detecting ∙ proactively block preemptively after financial for the first time deep learning skills' over banking applied to the transaction detection systems, the upgrade was a fraud prevention system.

To process large amounts of data through a neural network similar to the 'deep learning' is the human brain, and a technique for learning a new pattern, the system itself. Through a deep learning it can guard against financial fraud that increasingly intelligent grasp more than a new form of financial transaction is expected to play a big role in the electronic financial accidents. In addition, it can also reduce errors by more than aware of the normal trading transactions is expected to be improved ease trade customers.

This system was introduced in progress with a pin-tech enterprise development program seuraep Future 03 companies 'InfiniBand trees' and the collaboration of the Shinhan Financial. Shinhan Bank, InfiniBand trees ', developed by deep learning based on the "over financial transactions detection solution' is 'GruDEEP' by adding to the existing system in parallel to the existing trading patterns Model and deep learning model" over financial transactions detection system "that It implemented.

Shinhan Bank official said, "showed one of the" over financial transactions Detection Systems' newly introduced pre-test results, existing systems compared to 56% enhanced sophisticated fraud detection hit rates were detected by fraud patterns did not detect the existing system "and" new forms to effectively prevent fraud nagagetda said, "make a secure financial environment that allows customers to believe that changes in trading.





Financial Daily Economic News FNTIMES - unauthorized reproduction for commercial purposes in accordance with the Copyright Act, No copying, distribution,
Copyright ⓒ Korea's financial newspaper & FNTIMES.com