Fraud detection system, fraud detection method, and program
A fraud detection system including at least one processor configured to: calculate a fraud level of a user who uses a service based on a behavior of the user; obtain a determination result as to whether the behavior is actually fraudulent; and calculate accuracy of fraud detection for each user based on the fraud level and the determination result.
1. A fraud detection system comprising at least one processor configured to:
calculate a fraud level of a user who uses a service based on a behavior of the user obtained during a networked-session on the service;
obtain a determination result as to whether the behavior is actually fraudulent;
calculate accuracy of fraud detection for each user based on the fraud level and the determination result;
calculate a fraud level of each behavior based on each of a plurality of behaviors,
obtain a determination result as to whether each behavior is actually fraudulent,
calculate the accuracy based on a fraud level and a determination result of each behavior; and
store the accuracy associated with the user.
2. The fraud detection system according to claim 1 , wherein
the at least one processor calculates the fraud level of a behavior for which whether it is fraudulent has not been determined based on a learning model that has learned a behavior for which whether it is fraudulent has been determined as indicated by the determination result.
3. The fraud detection system according to claim 1 , wherein
the at least one processor calculates a value of a predetermined loss function based on the fraud level and the determination result, and obtains the calculated value as the accuracy.
4. The fraud detection system according to claim 3 , wherein
a coefficient of the loss function is determined such that a false negative has a greater impact on the accuracy as compared to a false positive.
5. The fraud detection system according to claim 1 , wherein
the at least one processor calculates the fraud level on an assumption that a behavior of another user for which whether it is fraudulent has been determined is performed by the user,
the at least one processor obtains the determination result of the behavior of the another user, and
the accuracy is calculated based on the fraud level of the behavior that is assumed to be performed by the user and the determination result of the behavior of the another user.
6. The fraud detection system according to claim 1 , wherein
the at least one processor determines a weighting coefficient in calculating the fraud level based on the fraud level according to a behavior that is previously performed and a determination result as to whether the behavior is actually fraudulent, and calculates the fraud level based on the determined weighting coefficient.
7. The fraud detection system of claim 1 , wherein
the at least one processor calculates the fraud level further based on at least one of an authentication method of the user and a holder of a credit card of the user.
8. The fraud detection system according to claim 1 , wherein the at least one processor executes processing according to the accuracy.
9. The fraud detection system according to claim 8 , wherein
the processing is processing for changing a value of the calculated fraud level.
10. The fraud detection system according to claim 8 , wherein
the processing is processing for providing the user with a coupon relating to the service.
11. A fraud detection method comprising:
calculating a fraud level of a user who uses a service based on a behavior of the user obtained during a networked-session on the service;
obtaining a determination result as to whether the behavior is actually fraudulent;
calculating accuracy of fraud detection for each user based on the fraud level and the determination result;
calculating a fraud level of each behavior based on each of a plurality of behaviors,
obtaining a determination result as to whether each behavior is actually fraudulent,
calculating the accuracy based on a fraud level and a determination result of each behavior; and
storing the accuracy associated with the user.
12. A non-transitory computer-readable information storage medium for storing a program for causing a computer to:
calculate a fraud level of a user who uses a service based on a behavior of the user obtained during a networked-session on the service;
obtain a determination result as to whether the behavior is actually fraudulent;
calculate accuracy of fraud detection for each user based on the fraud level and the determination result;
calculate a fraud level of each behavior based on each of a plurality of behaviors,
obtain a determination result as to whether each behavior is actually fraudulent,
calculate the accuracy based on a fraud level and a determination result of each behavior; and
store the accuracy associated with the user.