Fraud detection

Special-purpose systems that use advanced learning algorithms, can detect patterns of fraudulent transactions with credit cards or other financial transactions, based on the transaction features.

A sample of past transactions is used as a Historical Data Base. Some of these transactions proved fraudulent, some others proved OK.

The system can discover the patterns that lead to fraudulent transactions.

As a result:

  • Managers can explicitly learn which are the conditions which lead to fraudulent transactions.
  • The system can act so as to prevent a fraudulent transaction to occur by rejecting it. Thus, the Bank or Other Financial Authority automatically stops any suspicious transactions.
  • Managers are relieved from the burden of auditing such large volumes of transactions and get summary and analytical information on those transactions.