Method and system for detecting fraudulent financial transaction

1. A fraudulent financial transaction detection method performed by a computer apparatus comprising processing circuitry, the fraudulent financial transaction detection method comprising:

collecting, by the processing circuitry, transaction details from a financial institution;

first classifying, by the processing circuitry, each of a plurality of users into one among a set of first risk groups for a first transaction action type, the plurality of users corresponding to the transaction details, and each respective user among the plurality of users being classified into a corresponding first risk group among the set of first risk groups;

second classifying, by the processing circuitry, each of the plurality of users into one among a set of second risk groups for a second transaction action type, each respective user among the plurality of users being classified into a corresponding second risk group among the set of second risk groups, the second transaction action type being different from the first transaction action type;

determining, by the processing circuitry, whether a first user among the plurality of users is in a high-risk group based on a sum of scores set for a first set of risk groups, the first set of risk groups including,

the corresponding first risk group into which the first user is classified, and

the corresponding second risk group into which the first user is classified; and

blocking, by the processing circuitry, a requested transaction in response to determining the first user is in the high-risk group.

2. The fraudulent financial transaction detection method of claim 1 , wherein each of the first transaction action type and the second transaction action type comprises one of Buy, Sell, Send, Receive, Deposit, Withdrawal, External Send, or External Receive.

3. The fraudulent financial transaction detection method of claim 1 , wherein the first classifying and the second classifying comprises inputting the transaction details to a machine learning module trained to classify users into the set of first risk groups and the set of second risk groups based on corresponding transaction details.

4. The fraudulent financial transaction detection method of claim 3 , wherein the machine learning module is trained to output a score for each respective set of risk groups among the set of first risk groups and the set of second risk groups into which each corresponding user among the plurality of users is classified.

5. The fraudulent financial transaction detection method of claim 1 , wherein the first classifying includes,

calculating an average value for each of a plurality of first transaction action items corresponding to the first transaction action type, and

classifying each of the plurality of users into the one among the set of first risk groups by clustering the plurality of users based on a first distance between a value of each of the plurality of first transaction action items and the average value of each of the plurality of first transaction action items, the plurality of first transaction action items corresponding to a plurality of first transaction actions triggered by the plurality of users; and the second classifying includes,

calculating an average value for each of a plurality of second transaction action items corresponding to the second transaction action type, and

classifying each of the plurality of users into the one among the set of second risk groups by clustering the plurality of users based on a second distance between a value of each of the plurality of second transaction action items and the average value of each of the plurality of second transaction action items, the plurality of second transaction action items corresponding to a plurality of second transaction actions triggered by the plurality of users.

6. The fraudulent financial transaction detection method of claim 5 , wherein

the first classifying includes calculating a score of each among the set of first risk groups using a distance formula based on the first distance and a first directivity; and

the second classifying includes calculating a score of each among the set of second risk groups using the distance formula based on the second distance and a second directivity.

7. The fraudulent financial transaction detection method of claim 1 , wherein

the first classifying includes classifying each of the plurality of users into the one among the set of first risk groups based on a value of each of a plurality of first transaction action items and a standard deviation of each of the plurality of first transaction action items, the plurality of first transaction action items corresponding to a plurality of first transaction actions triggered by the plurality of users; and

the second classifying includes classifying each of the plurality of users into the one among the set of second risk groups based on a value of each of a plurality of second transaction action items and a standard deviation of each of the plurality of second transaction action items, the plurality of second transaction action items corresponding to a plurality of second transaction actions triggered by the plurality of users.

8. The fraudulent financial transaction detection method of claim 7 , wherein

each among the set of first risk groups is associated with a respective portion of a first probability density function corresponding to the plurality of first transaction action items, each respective portion of the first probability density function being separated by the standard deviation of each of the plurality of first transaction action items;

the first classifying includes classifying each of the plurality of users into the one among the set of first risk groups based on the respective portions of the first probability density function into which the value of each of the plurality of first transaction action items corresponds;

each among the set of second risk groups is associated with a respective portion of a second probability density function corresponding to the plurality of second transaction action items, each respective portion of the second probability density function being separated by the standard deviation of each of the plurality of second transaction action items; and

the second classifying includes classifying each of the plurality of users into the one among the set of second risk groups based on the respective portions of the second probability density function into which the value of each of the plurality of second transaction action items corresponds.

9. The fraudulent financial transaction detection method of claim 1 , further comprising:

generating a signal indicating that the first user is in the high-risk group in response to determining the first user is in the high-risk group.

10. The fraudulent financial transaction detection method of claim 9 , wherein the blocking comprises causing the requested transaction to be blocked by sending the signal to the financial institution.

11. The fraudulent financial transaction detection method of claim 1 , further comprising:

third classifying, by the processing circuitry, each of the plurality of users into one among a set of third risk groups for a third transaction action type, each respective user among the plurality of users being classified into a corresponding third risk group among the set of third risk groups, the third transaction action type being different from the first transaction action type and the second transaction action type,

wherein the first set of risk groups includes,

the corresponding first risk group into which the first user is classified,

the corresponding second risk group into which the first user is classified, and

the corresponding third risk group into which the first user is classified.

12. The fraudulent financial transaction detection method of claim 1 , wherein the determining comprises determining that the first user is in the high-risk group based on the sum of scores exceeding a threshold value associated with the high-risk group.

13. The fraudulent financial transaction detection method of claim 1 , wherein

each first risk group among the set of first risk groups is associated with a different respective first score;

each second risk group among the set of second risk groups is associated with a different respective second score; and

the method further comprises calculating the sum of scores as the sum of,

the respective first score associated with the corresponding first risk group; and

the respective second score associated with the corresponding second risk group.

14. A non-transitory computer-readable record medium storing instructions that, when executed by at least one processor, cause the at least one processor to perform a fraudulent financial transaction detection method, the fraudulent financial transaction detection method comprising:

collecting, by the at least one processor, transaction details from a financial institution;

first classifying, by the at least one processor, each of a plurality of users into one among a set of first risk groups for a first transaction action type, the plurality of users corresponding to the transaction details, and each respective user among the plurality of users being classified into a corresponding first risk group among the set of first risk groups;

second classifying, by the at least one processor, each of the plurality of users into one among a set of second risk groups for a second transaction action type, each respective user among the plurality of users being classified into a corresponding second risk group among the set of second risk groups, the second transaction action type being different from the first transaction action type;

determining, by the at least one processor, whether a first user among the plurality of users is in a high-risk group based on a sum of scores set for a first set of risk groups, the first set of risk groups including,

the corresponding first risk group into which the first user is classified, and

the corresponding second risk group into which the first user is classified; and

blocking, by the at least one processor, a requested transaction in response to determining the first user is in the high-risk group.

15. A computer apparatus comprising:

processing circuitry configured to cause the computer apparatus to,

collect transaction details from a financial institution,

first classify each of a plurality of users into one among a set of first risk groups for a first transaction action type, the plurality of users corresponding to the transaction details, and each respective user among the plurality of users being classified into a corresponding first risk group among the set of first risk groups,

second classify each of the plurality of users into one among a set of second risk groups for a second transaction action type, each respective user among the plurality of users being classified into a corresponding second risk group among the set of second risk groups, the second transaction action type being different from the first transaction action type,

determine whether a first user among the plurality of users is in a high-risk group based on a sum of scores set for a first set of risk groups, the first set of risk groups including,

the corresponding first risk group into which the first user is classified, and

the corresponding second risk group into which the first user is classified, and

block a requested transaction in response to determining the first user is in the high-risk group.

16. The computer apparatus of claim 15 , wherein the processing circuitry is configured to cause the computer apparatus to perform the first classification and the second classification including inputting the transaction details to a machine learning module trained to classify users into the set of first risk groups and the set of second risk groups based on corresponding transaction details.

17. The computer apparatus of claim 16 , wherein the machine learning module is trained to output a score for each respective set of risk groups among the set of first risk groups and the set of second risk groups into which each corresponding user among the plurality of users is classified.

18. The computer apparatus of claim 15 , wherein the processing circuitry is configured to cause the computer apparatus to:

first classify each of the plurality of users into the one among the set of first risk groups including,

calculating an average value for each of a plurality of first transaction action items corresponding to the first transaction action type, and

classify each of the plurality of users into the one among the set of first risk groups by clustering the plurality of users based on a first distance between a value of each of the plurality of first transaction action items and the average value of each of the plurality of first transaction action items; and

second classify each of the plurality of users into the one among the set of second risk groups including,

calculating an average value for each of a plurality of second transaction action items corresponding to the second transaction action type, and

classify each of the plurality of users into the one among the set of second risk groups by clustering the plurality of users based on a second distance between a value of each of the plurality of second transaction action items and the average value of each of the plurality of second transaction action items.

19. The computer apparatus of claim 18 , wherein the processing circuitry is configured to cause the computer apparatus to:

first classify each of the plurality of users into the one among the set of first risk groups including calculating a score of each among the set of first risk groups using a distance formula based on the first distance and a first directivity; and

second classify each of the plurality of users into the one among the set of second risk groups including calculating a score of each among the set of second risk groups using the distance formula based on the second distance and a second directivity.

20. The computer apparatus of claim 15 , wherein the processing circuitry is configured to cause the computer apparatus to:

first classify each of the plurality of users into the one among the set of first risk groups based on a value of each of a plurality of first transaction action items and a standard deviation each of the plurality of first transaction action items, the plurality of first transaction action items corresponding to a plurality of first transaction actions triggered by the plurality of users; and

second classify each of the plurality of users into the one among the set of second risk groups based on a value of each of a plurality of second transaction action items and a standard deviation each of the plurality of second transaction action items, the plurality of second transaction action items corresponding to a plurality of second transaction actions triggered by the plurality of users.

21. The computer apparatus of claim 15 , wherein each of the first transaction action type and the second transaction action type comprises one of Buy, Sell, Send, Receive, Deposit, Withdrawal, External Send, or External Receive.