Data classification methods and apparatus for use with data fusion

• US 7,516,111 B2
• Filed: 03/28/2007
• Issued: 04/07/2009
• Est. Priority Date: 09/28/2004
• Status: Active Grant

• Current Assignee: Nielsen Company (US), LLC
• Original Assignee: The Nielsen Company LLC (MOOD Media Corporation)
• Inventors: McMillan, Francis Gavin, Samson, Jerome
• Primary Examiner(s): Hirl; Joseph P
• Application Number:
  US11/692,480
• Publication Number:
  US 20070198573A1
• Time in Patent Office: 741 Days
• Field of Search: None
• US Class Current: 706/45
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising

Abstract

Methods and apparatus for classifying data for use in data fusion processes are disclosed. An example method of classifying data identifies classification tree information and assigns values to each of a plurality of nodes associated with the classification tree information. Each of the values is indicative of a relationship between its respective node and one of a plurality of classes. The example method also assigns each node to one of the classes based on a comparison of at least two of the assigned values and classifies the data based on the assignment of the nodes to the classes.

Claims

1. A computer implemented method of classifying data associated with characteristics of persons within a population of persons, comprising:

assigning a plurality of values to each of a plurality of terminal nodes associated with a previously generated classification tree, wherein each of the terminal nodes corresponds to a class of persons within the population of persons, and wherein each of the values associated with a particular terminal node is generated based on the frequency of occurrence of one of the classes at that terminal node and the frequency of occurrence of that one of the classes within the population of persons;

reassigning at least one of the plurality of terminal nodes to a different one of the classes of persons based on a comparison of at least two of the assigned values to form a modified classification tree;

classifying data records within the data based on the modified classification tree so that each data record within the data is associated with only one of the classes of persons; and

displaying information associated with characteristics of persons based on the classified data records.

2. A method as defined in claim 1, further comprising retrieving values representative of the classification tree from a memory.

3. A method as defined in claim 1, wherein each of the plurality of values corresponds to only one of the classes.

4. A method as defined in claim 1, wherein the values are index values.

5. A method as defined in claim 1, wherein reassigning at least one of the terminal nodes to a different one of the classes based on the comparison of the at least two of the assigned values results in at least two of the terminal nodes being assigned to the same one of the classes.

6. A method as defined in claim 1, wherein reassigning at least one of the terminal nodes includes reassigning the at least one of the terminal nodes to the different one of the classes of persons so that relatively larger assigned values determine the class to which that at least one of the terminal nodes is reassigned.

7. A method as defined in claim 1, wherein the classes of persons are usage or consumption classes.

8. A method as defined in claim 1, further comprising fusing the classified data records based on classification information associated with the classified data records.

9. An apparatus for classifying data associated with characteristics of persons within a population of persons, comprising:

a memory; and

a processor coupled to the memory and configured to;

assign a plurality of values to each of a plurality of terminal nodes associated with a previously generated classification tree, wherein each of the terminal nodes corresponds to a class of persons within the population of persons, and wherein each of the values associated with a particular terminal node is generated based on the frequency of occurrence of that one of the classes within the population of persons;

reassign at least one of the plurality of terminal nodes to a different one of the classes of persons based on a comparison of at least two of the assigned values to form a modified classification tree;

classify data records within the data based on the modified classification tree so that each data record within the data is associated with only one of the classes of persons; and

display information associated with characteristics of persons based on the classified data records.

10. An apparatus as defined in claim 9, wherein the processor is configured to retrieve values representative of the classification tree from a memory.

11. An apparatus as defined in claim 9, wherein each of the plurality of values corresponds to only one of the classes.

12. An apparatus as defined in claim 9, wherein the values are index values.

13. An apparatus as defined in claim 9, wherein the processor is configured to reassign at least one of the terminal nodes to a different one of the classes based on the comparison of the at least two of the assigned values so that at least two of the terminal nodes are assigned to the same one of the classes.

14. An apparatus as defined in claim 9, wherein the processor is configured to reassign at least one of the terminal nodes by reassigning at least one of the terminal nodes to the different one of the classes of persons so that relatively larger assigned values determine the class to which the at least one of the terminal nodes is reassigned.

15. An apparatus as defined in claim 9, wherein the classes of persons are usage or consumption classes.

16. An apparatus as defined in claim 9, wherein the processor is configured to fuse the classified data records based on classification information associated with the classified data records.

17. A machine accessible medium having instructions stored thereon that, when executed, cause a machine to:

assign a plurality of values to each of a plurality of terminal nodes associated with a previously generated classification tree, wherein each of the terminal nodes corresponds to a class of persons within a population of persons, and wherein each of the values associated with a particular terminal node is generated based on the frequency of occurrence of that one of the classes within the population of persons;

reassign at least one of the plurality of terminal nodes to a different one of the classes of persons based on a comparison of at least two of the assigned values to form a modified classification tree;

classify data records based on the modified classification tree so that each data record is associated with only one of the classes of persons; and

display information associated with characteristics of persons based on the classified data records.