Process and system for integrating information from disparate databases for purposes of predicting consumer behavior

• US 8,200,525 B2
• Filed: 01/02/2009
• Issued: 06/12/2012
• Est. Priority Date: 12/30/1998
• Status: Expired due to Fees

• Current Assignee: GfK US MRI, LLC
• Original Assignee: Experian Marketing Solutions Incorporated (Experian plc)
• Inventors: Kilger, Max F., Engel, William E.
• Primary Examiner(s): Sterrett, Jonathan G
• Application Number:
  US12/348,040
• Publication Number:
  US 20090182625A1
• Time in Patent Office: 1,257 Days
• Field of Search: 705/10, 705/7.31, 705/7.32, 705/7.33
• US Class Current: 705/7.31
• CPC Class Codes: G06F 16/27 Replication, distribution or synchronisation of data between databases or within a distributed database system; Distributed database system architectures therefor
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0202 Market predictions or forecasting for commercial activities
• CPC Class Codes: G06Q 30/0203 Market surveys; Market polls
• CPC Class Codes: G06Q 30/0204 Market segmentation
• CPC Class Codes: Y10S 707/99932 Access augmentation or optimizing
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access

Abstract

A process and system for integrating information stored in at least two disparate databases. The stored information includes consumer transactional information. According to the process and system, at least one qualitative variable which is common to each database is identified, and then transformed into one or more quantitative variables. The consumer transactional information in each database is then converted into converted information in terms of the quantitative variables. Thereafter, an integrated database is formed for predicting consumer behavior by combining the converted information from the disparate databases.

Claims

1. A computer-implemented method for predicting consumer behavior from information stored in a plurality of separate databases, the computer-implemented method comprising:

using a processing device, accessing a first database of the plurality of separate databases, the first database comprising a first plurality of records respectively associated with a first plurality of consumers, and the first plurality of records comprising a first plurality of data variables associated with survey data obtained from the first plurality of consumers;

using a processing device, accessing a second database of the plurality of separate databases, the second database comprising a second plurality of records respectively associated with a second plurality of consumers, the second plurality of records comprising a second plurality of data variables associated with transactions of the second plurality of consumers, wherein the first plurality of data variables includes data variables that are not included in the second plurality of data variables, and wherein at least some consumers of the second plurality of consumers are not in the first plurality of consumers;

identifying at least one qualitative data variable which is in both the first plurality of data variables and the second plurality of data variables;

transforming the at least one qualitative data variable into a plurality of quantitative variables;

using a processing device, differentially weighting the plurality of quantitative variables;

using a processing device, converting at least some of the first plurality of records and the second plurality of records according to the differentially weighted plurality of quantitative variables to form converted information;

using a processing device and using the converted information, performing a cluster analysis across the first and second databases to form a plurality of clusters, at least one cluster of the plurality of clusters containing at least some individuals of the first and second databases that are not in both the first and second databases;

using a processing device, linking through the plurality of clusters the first and second databases to form an integrated data structure;

using a processing device, for a selected cluster of the plurality of clusters, associating a plurality of additional behavioral characteristics of the consumers of the first plurality of consumers associated with the selected cluster, wherein the plurality of additional behavioral characteristics are not included in the qualitative and quantitative variables; and

using a processing device, predicting consumer behavior for the selected cluster using corresponding data of the selected cluster of the integrated data structure, the corresponding data including the plurality of additional behavioral characteristics of the selected cluster.

2. The computer-implemented method of claim 1, further comprising:

using a processing device, selecting at least one discriminating subset of the plurality of quantitative variables to create one or more statistical drivers; and

using a processing device, evaluating the pluralities of consumers represented in the plurality of separate databases using the one or more statistical drivers.

3. The computer-implemented method of claim 2, further comprising:

using a processing device, determining the one or more statistical drivers using a root mean squared standard statistic.

4. The computer-implemented method of claim 2, wherein the selecting step further comprises:

identifying one or more industries which have discriminating consumers and grouping selected commercial entities into the at least one discriminating subset.

5. The computer-implemented method of claim 1, further comprising:

converting one or more clusters of the plurality of clusters into at least one super cluster; and

assigning the consumers of the first and second pluralities of consumers to a corresponding cluster or super cluster using data from each database of the plurality of separate databases.

6. The computer-implemented method of claim 1, wherein the at least one qualitative variable is a commercial entity and the plurality of quantitative variables comprises two or more of the quantitative variables selected from the group consisting of:

mean number of transactions per person for the commercial entity, mean amount per transaction for the commercial entity, mean household income of consumers purchasing from the commercial entity, mean proportion of consumers for a particular geographic region of the commercial entity, mean proportion of consumers for a plurality of counties having different population sizes of a particular geographic region of the commercial entity, and combinations thereof.

7. The computer-implemented method of claim 1, wherein the step of differentially weighting the plurality of quantitative variables further comprises:

differentially weighting the plurality of quantitative variables to adjust for differences in numbers of transactions and in time periods encompassed.

8. The computer-implemented method of claim 1, wherein the information stored in at least the second database further comprises instances of purchasing behavior by consumers.

9. The computer-implemented method of claim 8, wherein at least one of the separate databases includes joint account information for at least two consumers, and wherein the method further comprises:

determining a consumer of the at least two consumers who generated at least a portion of the consumer transactional information.

10. The method of claim 1, wherein the step of predicting consumer behavior further comprises:

using a processing device, predicting a behavior of one or more consumers of the second plurality of consumers associated with a selected cluster of the plurality of clusters based at least in part on the plurality of additional behavioral characteristics of the consumers of the first plurality of consumers associated with the selected cluster.

11. A computer system for predicting consumer behavior from information stored in a plurality of separate databases, the computer system comprising:

one or more data storage devices to store a first database of the plurality of separate databases and a second database of the plurality of separate databases, the first database comprising a first plurality of records respectively associated with a first plurality of consumers, and the first plurality of records comprising a first plurality of data variables associated with survey data obtained from the first plurality of consumers;

the second database comprising a second plurality of records respectively associated with a second plurality of consumers, the second plurality of records comprising a second plurality of data variables associated with transactions of the second plurality of consumers, wherein the first plurality of data variables includes data variables that are not included in the second plurality of data variables, and wherein at least some consumers of the second plurality of consumers are not in the first plurality of consumers; and

one or more processing devices coupled to the one or more data storage devices, the one or more processing devices to identify at least one qualitative data variable which is in both the first plurality of data variables and the second plurality of data variables;

transform the at least one qualitative data variable into a plurality of quantitative variables;

differentially weight the plurality of quantitative variables;

convert at least some of the first plurality of records and the second plurality of records according to the differentially weighted plurality of quantitative variables to form converted information;

using the converted information, perform a cluster analysis across the first and second databases to form a plurality of clusters, at least one cluster of the plurality of clusters containing at least some individuals of the first and second databases that are not in both the first and second databases;

link through the plurality of clusters the first and second databases to form an integrated data structure and store the integrated data structure in the one or more data storage devices;

for a selected cluster of the plurality of clusters, associate a plurality of additional behavioral characteristics of the consumers of the first plurality of consumers associated with the selected cluster of the plurality of clusters, wherein the plurality of additional behavioral characteristics are not included in the qualitative and quantitative variables; and

predict consumer behavior for the selected cluster using corresponding data of the selected cluster of the integrated data structure, the corresponding data including the plurality of additional behavioral characteristics of the selected cluster.

12. The computer system of claim 11, wherein the one or more processing devices further is to select at least one discriminating subset of the plurality of quantitative variables to create one or more statistical drivers;

and evaluate the pluralities of consumers represented in the plurality of separate databases using the one or more statistical drivers.

13. The computer system of claim 12, wherein the one or more processing devices further is to determine the one or more statistical drivers using a root mean squared standard statistic.

14. The computer system of claim 12, wherein the one or more processing devices further is to identify one or more industries which have discriminating consumers and group selected commercial entities into the at least one discriminating subset.

15. The computer system of claim 11, wherein the one or more processing devices further is to convert one or more clusters of the plurality of clusters into at least one super cluster;

and assign the consumers of the first and second pluralities of consumers to a corresponding cluster or super cluster using data from each database of the plurality of separate databases.

16. The computer system of claim 11, wherein the at least one qualitative variable is a commercial entity and the plurality of quantitative variables comprises two or more of the quantitative variables selected from the group consisting of:

mean number of transactions per person for the commercial entity, mean amount per transaction for the commercial entity, mean household income of consumers purchasing from the commercial entity, mean proportion of consumers for a particular geographic region of the commercial entity, mean proportion of consumers for a plurality of counties having different population sizes of a particular geographic region of the commercial entity, and combinations thereof.

17. The computer system of claim 11, wherein the one or more processing devices further is to differentially weight the plurality of quantitative variables to adjust for differences in numbers of transactions and in time periods encompassed.

18. The computer system of claim 11, wherein the information stored in at least the second database further comprises instances of purchasing behavior by consumers.

19. The computer system of claim 18, wherein at least one of the separate databases includes joint account information for at least two consumers, and wherein the one or more processing devices further is to determine a consumer of the at least two consumers who generated at least a portion of the consumer transactional information.

20. The computer system of claim 11, wherein the one or more processing devices further is to predict a behavior of one or more consumers of the second plurality of consumers associated with a selected cluster of the plurality of clusters based at least in part on the plurality of additional behavioral characteristics of the consumers of the first plurality of consumers associated with the selected cluster.

21. The computer system of claim 11, wherein the information stored in the plurality of separate databases further comprises one or more of types of information selected from the group consisting of:

behavioral and attitudinal information;

consumer transactional information;

media consumption information;

instances of purchasing behavior by consumers;

survey information;

self-reported behavioral information; and

combinations thereof.

22. The computer system of claim 11, wherein the one or more processing devices further is to perform a principal components analysis on the plurality of quantitative variables using the information stored in each separate database of the plurality of separate and disparate databases to create a plurality of statistical drivers, and to convert the information stored in the plurality of separate databases according to the plurality of statistical drivers to create the converted information.

23. The computer system of claim 22, wherein the one or more processing devices further is to validate the plurality of clusters as a discriminatory behavioral model for predicting consumer behavior.

24. The computer system of claim 23, wherein the one or more processing devices further is to determine whether the plurality of clusters provides corresponding discrimination on a plurality of other variables which are not statistical drivers in the plurality of separate databases.

25. The computer system of claim 23, wherein the one or more processing devices further is to evaluate corresponding discrimination power of the plurality of statistical drivers using a second cluster analysis.

26. The computer system of claim 11, wherein the one or more processing devices further is to standardize the plurality of quantitative variables and transform the standardized plurality of quantitative variables to be substantially orthogonal.