System architecture for fraud detection

1. A system for performing analysis preprocessing for detecting fraud, comprising

a memory comprising executable instructions;

a processor configured to execute the executable instructions and cause the system to:

receive event data from a user at a client system, wherein the event data comprising a channel identifier and a token;

determine if the token is valid for the client system;

verify validity of the event data by checking format of the event data against a standard format upon determination that the token is valid;

upon verification of validity of the event data, augment the event data with additional context data, the additional context data including a frequency of contact with the client system by the channel identifier;

determine that the identifier is not on a whitelist;

and

in response to determining that the channel identifier is not on the whitelist,

for each behavior factor of a plurality of behavior factors associated with the channel identifier, calculate a mean for the behavior factor;

for each behavior factor of the plurality of behavior factors, determine a deviation from the calculated mean for the behavior factor based on the augmented event data; and

score the augmented event data based on possible attack vectors based on the determined deviations for each of the plurality of behavioral factors associated with the channel identifier, based on situational factors associated with a channel address, a call origination location, and a service provider; and forward the augmented event data for analysis.

2. The system of claim 1 , wherein the client system comprises an application program interface and the token is an API token.

3. The system of claim 1 , wherein to score the augmented event data, the executable instructions, when executed, further cause the system to check of how many times a user has contacted the system and how many accounts the user has attempted to access over a predetermined period of time.

4. The system of claim 1 , wherein to score the augmented event data, the executable instructions, when executed, further cause the system to check the augmented event data against criteria to determine a risk associated with the event data.

5. The system of claim 4 , wherein the criteria comprises behavior criterion.

6. The system of claim 4 , wherein the criteria comprises situation criterion.

7. The system of claim 4 , wherein the criteria comprises reputation criterion.

8. The system of claim 1 , wherein to score the augmented event data, the executable instructions, when executed, further cause the system to compare the augmented event data against an expected pattern of normal use.

9. A non-transitory computer readable medium comprising instructions that, when executed by a processor of a processing system, cause the processing system to perform a method, the method comprising:

receive event data from a user at a client system, wherein the event data comprising a channel identifier and a token;

determine if the token is valid for the client system;

verify validity of the event data by checking format of the event data against a standard format upon determination that the token;

upon verification of validity of the event data, augment the event data with additional context data, the additional context data including a frequency of contact with the client system by the channel identifier;

determine that the identifier is not on a whitelist;

and

in response to determining that the channel identifier is not on the whitelist,

for each behavior factor of a plurality of behavior factors associated with the channel identifier, calculate a mean for the behavior factor;

for each behavior factor of the plurality of behavior factors, determine a deviation from the calculated mean for the behavior factor based on the augmented event data; and

score the augmented event data based on possible attack vectors based on the determined deviations for each of the plurality of behavioral factors associated with the channel identifier, based on situational factors associated with a channel address, a call origination location, and a service provider; and forward the augmented event data for analysis.

10. The non-transitory computer readable medium of claim 9 , wherein the client system comprises an application program interface and the token is an API token.

11. The non-transitory computer readable medium of claim 9 , wherein to score the augmented event data, the executable instructions, when executed, further cause the system to check of how many times a user has contacted the system and how many accounts the user has attempted to access over a predetermined period of time.

12. The non-transitory computer readable medium of claim 9 , wherein to score the augmented event data, the executable instructions, when executed, further cause the system to check the augmented event data against criteria to determine a risk associated with the event data.

13. The non-transitory computer readable medium of claim 12 , wherein the criteria comprises behavior criterion.

14. The non-transitory computer readable medium of claim 12 , wherein the criteria comprises situation criterion.

15. The non-transitory computer readable medium of claim 12 , wherein the criteria comprises reputation criterion.

16. The non-transitory computer readable medium of claim 9 , wherein to score the augmented event data, the executable instructions, when executed, further cause the system to compare the augmented event data against an expected pattern of normal use.

17. A method, comprising:

receiving event data from a user at a system, wherein the event data comprising a channel identifier and a token;

determining if the token is valid for the system;

verifying validity of the event data by checking format of the event data against a standard format upon determination that the token is valid;

upon verification of validity of the event data, augmenting the event data with additional context data, the additional context data including a frequency of contact with the system by the channel identifier;

determining that the identifier is not on a whitelist;

and

in response to determining that the channel identifier is not on the whitelist,

for each behavior factor of a plurality of behavior factors associated with the channel identifier, calculating a mean for the behavior factor;

for each behavior factor of the plurality of behavior factors, determining a deviation from the calculated mean for the behavior factor based on the augmented event data; and

scoring the augmented event data based on possible attack vectors based on the determined deviations for each of the plurality of behavioral factors associated with the channel identifier, based on situational factors associated with a channel address, a call origination location, or a service provider; and forwarding the augmented event data for analysis.

18. The method of claim 17 , wherein the system comprises an application program interface and the token is an API token.

19. The method of claim 17 , wherein the scoring is threshold check of how many times a user has contacted the system and how many accounts the user has attempted to access over a predetermined period of time.

20. The method of claim 17 , wherein the scoring comprises checking the augmented event data against criteria to determine a risk associated with the event data.

21. The method of claim 20 , wherein the criteria comprises behavior criterion.

22. The method of claim 20 , wherein the criteria comprises situation criterion.

23. The method of claim 20 , wherein the criteria comprises reputation criterion.

24. The method of claim 17 , wherein the scoring comprises comparing the augmented event data against an expected pattern of normal use.

25. The method of claim 17 , further comprising partitioning the recorded event data using a unique hash key.

26. The method of claim 17 , wherein the behavioral factors are determined by calculating a statistical standard deviation over a predetermined period of time for a typical contact with the client system and determining whether the event data falls outside the statistical deviation.

27. The non-transitory computer readable medium of claim 9 , the method further comprising partitioning the recorded event data using a unique hash key.