OLAP-based customer behavior profiling method and system

• US 8,190,554 B2
• Filed: 07/01/2004
• Issued: 05/29/2012
• Est. Priority Date: 03/10/2000
• Status: Active Grant

• Current Assignee: Valtrus Innovations Limited (f/k/a Dolya Holdco 9 Limited) (Key Patent Innovations Limited)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Chen, Qiming, Dayal, Umeshwar, Hsu, Meichun
• Primary Examiner(s): Wu, Yicun
• Application Number:
  US10/882,925
• Publication Number:
  US 20040249866A1
• Time in Patent Office: 2,889 Days
• Field of Search: 707 1- 10, 707100-1041, 707200-206
• US Class Current: 707/600
• CPC Class Codes: G06F 16/283 Multi-dimensional databases or data warehouses, e.g. MOLAP or ROLAP
• CPC Class Codes: G06Q 40/08 Insurance
• CPC Class Codes: H04M 15/00 Arrangements for metering, time-control or time indication ; Metering, charging or billing arrangements for voice wireline or wireless communications, e.g. VoIP
• CPC Class Codes: H04M 15/04 Recording calls , or communications in printed, perforated or other permanent form
• CPC Class Codes: H04M 15/43 Billing software details
• CPC Class Codes: H04M 15/47 Fraud detection or prevention means
• CPC Class Codes: H04M 15/58 based on statistics of usage or network monitoring
• CPC Class Codes: H04M 2203/6027 Fraud preventions
• CPC Class Codes: H04M 2215/0148 Fraud detection or prevention means
• CPC Class Codes: H04M 2215/0188 Network monitoring; statistics on usage on called/calling number
• CPC Class Codes: H04M 3/2218 Call detail recording
• CPC Class Codes: Y10S 707/99943 Generating database or data structure, e.g. via user interface

Abstract

An OLAP-based method and system for profiling customer behavior that can be utilized to detect telecommunication fraud. First, call records are received. Next, a calling profile cube (e.g., a multi-customer profile cube) is generated based on the call records. A volume-based calling pattern cube (e.g., a calling pattern cube for each individual customer) is then generated based on the multi-customer profile cube. The volume-based calling pattern cube is then compared with known fraudulent volume-based calling patterns. If the similarities generated by the comparison reaches or exceeds a predetermined threshold, then the particular caller with the calling pattern being analyzed is considered suspicious. In this manner, suspicious calling activity can be detected, and appropriate remedial actions, such as further investigation or the cancellation of telephone services, can be taken.

Claims

1. A data processing system comprising:

a data warehouse for storing data in a relational format, said data warehouse including a profile table and a call table having multiple dimensions;

an OLAP server, coupled to the data warehouse, for providing predetermined OLAP operations that convert the profile table to a multi-dimensional profile cube and the call table to a multi-dimensional calling cube, update the profile cube with the multi-dimensional calling cube, generate individual caller pattern cubes from the updated profile cube, and store the updated profile cube in the data warehouse in the relational format, wherein the multi-dimensional calling cube has multiple levels,

the OLAP server to further utilize the caller pattern cubes to detect telecommunication fraud.

2. A data processing system as in claim 1 further comprising:

an analysis tool for use by a data analyst to perform one of comparing calling pattern cubes to known fraudulent calling pattern cubes and extracting information from the calling pattern cubes based on selected dimensions, levels, and ad-hoc queries provided by the data analyst.

3. A data processing system as in claim 1 further comprising:

a visualization tool for use by a data analyst to display the calling pattern cubes in different formats, levels, and dimensions.

4. A data processing system as in claim 1 wherein the OLAP server comprises:

a data staging tool for transferring data between the updated profile cube and the profile table stored in the data warehouse at predetermined time intervals.

5. The system of claim 1 wherein the profile cube contains information relating to telephone calls, and the caller pattern cubes represent calling behavior of individual callers.

6. A system, comprising:

a data warehouse that stores call records in a non-cube format; and

a computer system coupled to the data warehouse, the computer system receives the call records and maintains a caller profile cube for use with detecting telecommunication fraud;

wherein the caller profile cube comprises a combination of a profile cube and a snapshot profile cube, wherein the caller profile cube has multiple dimensions and multiple levels, and

wherein information associated with the caller profile cube is stored in the data warehouse in the non-cube format.

7. The system of claim 6 wherein the caller profile cube is periodically updated by converting the information associated with the caller profile cube from the non-cube format to a cube format and combining the caller profile cube with a new snapshot profile cube generated from new call records.

8. The system of claim 6 wherein the computer system uses OLAP programming to maintain the caller profile cube.

9. The system of claim 6 wherein the computer system uses calling parameters to generate probability-based calling pattern cubes for use with detecting telecommunication fraud.

10. The system of claim 9 wherein the calling parameters comprise at least one of a call duration, a time of day, a day of week, and a number of calls.

11. The system of claim 9 wherein the call records comprise telephone call records, the caller profile cube contains information relating to telephone calls, the snapshot profile cube is populated with information from the telephone call records, and the caller pattern cubes represent calling behavior of individual callers.

12. The system of claim 6 wherein the call records comprise telephone call records, the caller profile cube contains information relating to telephone calls, and the snapshot profile cube is populated with information from the telephone call records.

13. A system, comprising:

a data warehouse configured to store a behavioral profile table having multiple dimensions in a non-cube format;

a computer system configured to receive the behavioral profile table and convert the behavioral profile table into a behavioral profile cube that has multiple dimensions and multiple levels,

wherein the computer system is further configured to update the behavioral profile cube and convert the updated behavioral profile cube into the non-cube format for storage in the data warehouse.

14. The system of claim 13 wherein the computer system updates the behavioral profile cube by converting update data in a non-cube format to an update data cube having the same dimensions as the behavioral profile cube and combining the behavioral profile cube and the update data cube.

15. The system of claim 14 wherein the behavioral profile cube is associated with multiple consumers.

16. The system of claim 15 wherein the computer system is further configured to generate individual behavioral profile cubes from the updated behavioral profile cube.

17. The system of claim 16 wherein the individual behavioral profile cubes are used to detect fraud.

18. The system of claim 16 wherein the updated behavioral cube contains information relating to telephone calls, and the individual behavioral profile cubes represent telephone calling behavior of individual callers.

19. A method for detecting telecommunication fraud performed in a data processing system having a data warehouse and an OLAP server, the method comprising:

retrieving a plurality of call records from the data warehouse;

generating a calling profile cube based on the call records, wherein the calling profile cube includes information on multiple customers;

converting the calling profile cube into a non-cube format for storage in a data warehouse;

generating a volume-based calling pattern cube for each individual customer based on the multi-customer calling profile cube;

comparing the volume-based calling pattern cube for each customer to a predetermined fraudulent volume-based calling pattern; and

when the volume-based calling pattern cube is in a first predetermined relationship with predetermined fraudulent volume-based calling pattern, performing a first action.

20. A method as in claim 19 further comprising:

analyzing the calling pattern cube by utilizing at least one OLAP operation.

21. A method as in claim 20 wherein said OLAP operations is one of a roll-up operation, a drill-down operation, a dice operation, a slice operation, and an ad-hoc query.

22. A method as in claim 19 wherein the predetermined fraudulent volume-based calling pattern is one of a personalized calling pattern and a group-based pattern.

23. A method as in claim 19 further comprising:

utilizing an OLAP server to generate the calling profile cube, update the calling profile cube, derive calling pattern cubes from the updated calling profile cube, analyze the calling pattern cubes, and compare calling pattern cubes;

wherein OLAP programming supported by the OLAP server provides a scalable computation engine for generating and processing the calling pattern cubes.

24. The method of claim 19 wherein generating the calling profile cube comprises generating the calling profile cube containing information relating to telephone calls, and wherein generating the calling pattern cube comprises generating the calling pattern cube that represents calling behavior of an individual customer.

25. A method for detecting telecommunication fraud performed in a data processing system having a data warehouse and an OLAP server, the method comprising:

retrieving a plurality of call records from the data warehouse;

generating a calling profile cube based on the call records, wherein the calling profile cube includes information on multiple customers;

converting the calling profile cube into a non-cube format for storage in a data warehouse;

generating a volume-based calling pattern cube for each individual customer based on the multi-customer calling profile cube;

generating a probability-based calling pattern cube based on the volume-based calling pattern cube for each individual customer;

comparing the probability-based calling pattern cube for each customer to a predetermined fraudulent probability-based calling pattern;

when the probability-based calling pattern cube is in a first predetermined relationship with predetermined fraudulent probability-based calling pattern, performing a first action.

26. The method of claim 25 wherein the calling profile cube is a multi-dimensional and a multi-level cube and wherein the volume-based calling pattern cubes are multi-dimensional and a multi-level cubes.

27. The method of claim 26 wherein the dimensions include a day-of-week hierarchy, a time hierarchy, and a duration hierarchy.

28. The method of claim 25 further comprising:

performing data staging at predetermined time intervals; and

updating the calling profile cube by generating a snapshot cube from a call table; and

merging the snapshot cube with the calling profile cube to generate an updated calling profile cube.

29. The method of claim 25 wherein the calling profile cube has a cell that includes a probability distribution value based on one of the probability distribution on calls to each callee and the probability distribution on all calls.

30. The method of claim 25 wherein generating the calling profile cube comprises generating the calling profile cube containing information relating to telephone calls, and wherein generating the probability-based calling pattern cube comprises generating the probability-based calling pattern cube that represents calling behavior of an individual customer.

31. A method, comprising:

maintaining a behavioral profile table having multiple dimensions in a non-cube format;

converting the behavioral profile table into a behavioral profile cube that has multiple dimensions and multiple levels;

updating the behavioral profile cube;

converting the updated behavioral profile cube into the non-cube format.

32. The method of claim 31 wherein updating the behavioral profile cube comprises:

converting update data from a non-cube format to an update data cube having the same dimensions as the behavioral profile cube; and

combining the behavioral profile cube and the update data cube.

33. The method of claim 31 further comprising generating a plurality of individual behavioral profile cubes from the updated behavioral profile cube.

34. The method of claim 33 further comprising analyzing the individual behavioral profile cubes to detect fraud.

35. The method of claim 33 wherein the updated behavioral cube contains information relating to telephone calls, and wherein generating the individual behavioral cubes comprises generating the individual behavioral profile cubes that represent telephone calling behavior of individual callers.