The neural network created by Russian scientists can automatically detect suspicious financial transactions to combat fraud. In this respect informs TASS.
Development, Peter the Great St. Petersburg Polytechnic University (SPbPU). The system works on the principle of deep learning and takes into account parameters such as the type and number of the bank card, information about its owner, data about the recipient of the funds, how the transaction was carried out (through a banking application or otherwise). ) as well as reveal subtle patterns in past transmissions and transactions.
“If a person opened a bank account six months ago and the average daily transaction amount was 1,000 rubles during that time, and then received a money transfer of 30,000 rubles in one day, the probability of having a neural network will increase, classifying that person as a fraud,” the scientists explain.
During the tests, the system showed high efficiency and was found suitable for practical use. According to the creators, the neural network could be the first loop to protect bank customers and serve to detect illegal actions.
ancient scientists clarified a new type of activity in the brain of a dying person.