Computer-implemented data storage systems and methods for use with predictive model systems

• US 7,912,773 B1
• Filed: 03/26/2007
• Issued: 03/22/2011
• Est. Priority Date: 03/24/2006
• Status: Active Grant

• Current Assignee: SAS Institute Incorporated
• Original Assignee: SAS Institute Incorporated
• Inventors: Dulany, Paul C., Drossu, Radu, Subramanian, Revathi, Chen, Kevin Chaowen
• Primary Examiner(s): Kramer; James A
• Assistant Examiner(s): Gottschalk; Martin A
• Application Number:
  US11/691,277
• Time in Patent Office: 1,457 Days
• Field of Search: 375/E7.199, 375/240, 375/240.12, 235/375, 235/379, 706/47, 705/35, 705/39, 705/43, 705/44, 705/75, 707/200
• US Class Current: 705/35
• CPC Class Codes: G06Q 10/04 Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem" market predictions or forecasting for commercial activities G06Q30/0202
• CPC Class Codes: G06Q 20/4016 involving fraud or risk level assessment in transaction processing
• CPC Class Codes: G06Q 30/0185 Product, service or business identity fraud
• CPC Class Codes: G06Q 30/0202 Market predictions or forecasting for commercial activities
• CPC Class Codes: G06Q 30/06 Buying, selling or leasing transactions
• CPC Class Codes: G06Q 40/00 Finance; Insurance; Tax strategies; Processing of corporate or income taxes
• CPC Class Codes: G06Q 40/02 Banking, e.g. interest calculation or account maintenance credit or loans G06Q40/03
• CPC Class Codes: G06Q 40/03 Credit; Loans; Processing thereof
• CPC Class Codes: G06Q 40/06 Asset management; Financial planning or analysis
• CPC Class Codes: G06Q 40/12 Accounting

Abstract

Systems and methods for performing fraud detection. As an example, a system and method can be configured to contain a raw data repository for storing raw data related to financial transactions. A data store contains rules to indicate how many generations or to indicate a time period within which data items are to be stored in the raw data repository. Data items stored in the raw data repository are then accessed by a predictive model in order to perform fraud detection.

Claims

1. A computer-implemented system for performing fraud detection, comprising:

one or more processors;

a computer-readable storage medium containing instructions operable to cause the one or more processors to perform operations, including;

storing a set of raw data in a raw data repository including unprocessed financial transaction data records resulting from one or more financial transactions, each unprocessed transaction data record having one or more associated data fields;

selecting a subset of the set of raw data, wherein a statistical analysis is performed on the subset of the set of raw data;

performing the same statistical analysis on the set of raw data in the raw data repository;

comparing the results of the statistical analysis on the set of raw data and the results of the statistical analysis on the subset of the set of raw data to determine whether the results satisfy a predetermined criteria, wherein when the results satisfy the predetermined criteria, a storage rule is generated, the storage rule indicating how many records of each data field associated with each unprocessed transaction data record are to be stored in the raw data repository, or a time period during which each data field associated with each unprocessed transaction data record is to be stored in the raw data repository; and

using a predictive model to perform fraud detection on the raw data stored according to the storage rule.

2. The system of claim 1, wherein the raw data includes authorization data, posting or settlement data, statement data, payment data, and non-monetary data that are to be used by the predictive model to perform fraud detection.

3. The system of claim 2, wherein the non-monetary data includes address change data, phone change data, mother name change data, or credit line change data.

4. The system of claim 1, wherein the raw data includes unprocessed data records as originally received from the locations where the financial transactions occurred.

5. The system of claim 1, wherein the raw data includes unaggregated and underived data records.

6. The system of claim 1, wherein the storage rule is stored in an IF-THEN-ELSE format.

7. The system of claim 1, wherein the storage rule indicates how many records of raw payment amounts should be stored.

8. The system of claim 1, wherein the raw data includes one or more data records, and wherein the storage rule indicates the number of records of each data field to store based upon type of transaction or number of fields associated with the one or more data records.

9. The system of claim 1, wherein the raw data includes one or more data records, and wherein the storage rule indicates a particular timeframe for determining how long a data record is to be stored in the raw data repository.

10. The system of claim 1, wherein the raw data includes one or more data records, wherein the one or more data records are stored in the raw data repository in variable lengths;

and wherein the length of a data record is based upon a type of transaction with which the data record is associated.

11. The system of claim 1, wherein the records of each data field are stored in the raw data repository using a circular list for each data field.

12. The system of claim 1, wherein the raw data includes one or more data records, and wherein the one or more data records are at a transaction level or at an individual field level.

13. The system of claim 1, wherein the predictive model is a neural network predictive model that is configured to perform the fraud detection.

14. The system of claim 1, wherein the stored raw data includes one record per account or multiple records per account.

15. The system of claim 1, wherein the fraud detection is performed to determine whether fraud has occurred with respect to an entity;

and wherein the entity is a financial transaction card, an individual, a merchant, or an account.

16. The system of claim 1, wherein the storage rule indicates that a different amount of raw data is to be stored for different types of entities.

17. The system of claim 1, wherein the stored raw data includes incremental transaction data, the system further comprising instructions operable to cause the one or more processors to perform operations, including:

generating a first score based upon the raw data regarding a new incremental transaction with respect to an entity;

receiving a trigger generated independent of the occurrence of the new incremental transaction; and

generating, in response to the generated trigger, a second score for the entity, the second score being based upon stored past incremental transaction data and upon non-incremental transaction data;

wherein the generated second score is indicative of whether fraud has occurred or not; and

wherein the second score is used to determine whether a fraud handling action is to be performed upon the entity.

18. The system of claim 1, wherein the raw data repository is configured for access by different types of financial analysis systems.

19. The system of claim 1, wherein the raw data repository is configured for access by a financial fraud detection system, a loan risk analysis system, and a revenue generation prediction system.

25. The system of claim 1, wherein when comparing the results of the statistical analysis on the set of raw data and the results of the statistical analysis on the subset of the set of raw data, the results do not satisfy the predetermined criteria, then:

a second subset of the set of raw data is selected,

a statistical analysis is performed on the second subset of the set of raw data,

the results of the statistical analysis on the second subset of the set of raw data are compared to the results of the statistical analysis on the set of raw data, and

a storage rule is generated when the comparison associated with the statistical analysis on the second subset of the set of raw data indicates that the results satisfy the predetermined criteria.

20. A computer-implemented method for performing fraud detection, comprising:

storing, using one or more processors, a set of raw data in a raw data repository including unprocessed financial transaction data records resulting from one or more financial transactions, each unprocessed financial transaction data record having one or more associated data fields;

selecting a subset of the set of raw data, wherein a statistical analysis is performed on the subset of the set of raw data;

performing the same statistical analysis on the set of raw data in the raw data repository;

comparing the results of the statistical analysis on the set of raw data and the results of the statistical analysis on the subset of the set of raw data to determine whether the results satisfy a predetermined criteria, wherein when the results satisfy the predetermined criteria, a storage rule is generated, the storage rule indicating how many records of each data field associated with each unprocessed transaction data record are to be stored in the raw data repository, or a time period during which each data field associated with each unprocessed transaction data record is to be stored in the raw data repository; and

using a predictive model to perform fraud detection on the raw data stored according to the storage rule.

21. A computer-readable storage medium encoded with instructions that cause a computer to perform a method for detecting fraud, said method comprising the steps of:

storing a set of raw data in a raw data repository including unprocessed financial transaction data records resulting from one or more financial transactions, each unprocessed financial transaction data record having one or more associated data fields;

selecting a subset of the set of raw data, wherein a statistical analysis is performed on the subset of the set of raw data;

performing the same statistical analysis on the set of raw data in the raw data repository;

comparing the results of the statistical analysis on the set of raw data and the results of the statistical analysis on the subset of the set of raw data to determine whether the results satisfy a predetermined criteria, wherein when the results satisfy the predetermined criteria, a storage rule is generated, the storage rule indicating how many records of each data field associated with each unprocessed transaction data record are to be stored in the raw data repository, or a time period during which each data field associated with each unprocessed transaction data record is to be stored in the raw data repository; and

using a predictive model to perform fraud detection on the raw data stored according to the storage rule.

22. A computer-implemented system for performing fraud detection, comprising:

one or more processors;

a computer-readable storage medium containing instructions operable to cause the one or more processors to perform operations, including;

storing a set of raw data in a raw data repository including unprocessed financial transaction data records resulting from one or more financial transactions, each unprocessed transaction data record having two or more associated data fields;

selecting a subset of set of the raw data, wherein a statistical analysis is performed on the subset of the set of raw data;

performing the same statistical analysis on the set of raw data in the raw data repository;

comparing the results of the statistical analysis on the set of raw data and the results of the statistical analysis on the subset of the set of raw data to determine whether the results satisfy a predetermined criteria, wherein when the results satisfy the predetermined criteria, a storage rule is generated, the storage rule indicating how many records of each data field associated with each unprocessed transaction data record are to be stored in the raw data repository, wherein a first number of records of a first data field are stored, and wherein a second different number of records of a second data field are stored; and

using a predictive model to perform fraud detection on the raw data stored according to the storage rule.

23. A computer-implemented system for performing fraud detection, comprising:

one or more processors;

a computer-readable storage medium containing instructions operable to cause the one or more processors to perform operations, including;

storing a set of raw data in a raw data repository including unprocessed financial transaction data records resulting from one or more financial transactions, each financial transaction being a particular type of transaction, each unprocessed transaction data record having one or more associated fields;

selecting a subset of set of the raw data, wherein a statistical analysis is performed on the subset of set of the raw data;

performing the same statistical analysis on the set of raw data in the raw data repository;

comparing the results of the statistical analysis on the set of raw data and the results of the statistical analysis on the subset of the set of raw data to determine whether the results satisfy a predetermined criteria, wherein when the results satisfy the predetermined criteria, a storage rule is generated, wherein the storage rule indicates the number of the financial transaction data records to store based upon type of transaction or number of fields associated with the one or more data records; and

using a predictive model to perform fraud detection on the raw data stored according to the storage rule.

24. A computer-implemented system for performing fraud detection, comprising:

one or more processors;

a computer-readable storage medium containing instructions operable to cause the one or more processors to perform operations, including;

storing a set of raw data in a raw data repository including unprocessed financial transaction data records resulting from one or more financial transactions, each unprocessed transaction data record having one or more associated data fields;

selecting a subset of set of the raw data, wherein a statistical analysis is performed on the subset of the set of raw data;

performing the same statistical analysis on the set of raw data in the raw data repository;

comparing the results of the statistical analysis on the set of raw data and the results of the statistical analysis on the subset of the set of raw data, wherein when the comparison meets a predetermined criteria, a storage rule is generated, the storage rule indicating how many records of each data field associated with each unprocessed transaction data record are to be stored in the raw data repository and a particular timeframe for determining how long each unprocessed transaction data record is to be stored in the raw data repository; and

using a predictive model to perform fraud detection on the raw data stored according to the storage rule.