SYSTEMS AND METHODS FOR BEHAVIORAL PATTERN MINING

• US 20130346447A1
• Filed: 06/21/2012
• Published: 12/26/2013
• Est. Priority Date: 06/21/2012
• Status: Active Grant

• Current Assignee: Conduent Business Services, LLC (Conduent, Inc.)
• Original Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Inventors: Gnanasambandam, Shanmuga-Nathan, Jung, Gueyoung, Zhao, Shi, Wu, Changjun
• Granted Patent: US 9,305,104 B2
• Time in Patent Office: Days
• US Class Current: 707/776
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06F 16/955 using information identifiers, e.g. uniform resource locators [URL]

Abstract

Methods and systems of performing data mining may include receiving a plurality of web log records and a plurality of call log records; associating one or more web log records with a call log record, wherein the associated user for each of the associated one or more web log records and the call log record are the same; identifying one or more patterns among the web log records for the plurality of call log records, wherein each pattern comprises one or more web accesses, a time stamp at which each of the one or more web accesses is performed and the call topic for the call log record; identifying one or more web log records associated with a new call, and predicting a call topic for the new call based on at least one pattern and the one or more web log records.

Claims

1. A method of performing data mining, the method comprising:

receiving a web log having a plurality of web log records, wherein each web log record comprises a web access, a time stamp and an associated user;

receiving a call log having a plurality of call log records, wherein each call log record comprises a call topic, a time stamp and an associated user;

for each call log record, associating, by a processing device, one or more web log records with the call log record, wherein the associated user for each of the associated one or more web log records and the call log record are the same;

identifying, by the processing device, one or more patterns among the web log records for the plurality of call log records, wherein each pattern comprises one or more web accesses, a time stamp at which each of the one or more web accesses is performed and the call topic for the call log record;

identifying, by the processing device, one or more web log records associated with a new call; and

predicting, by the processing device, a call topic for the new call based on at least one pattern and the one or more web log records.

2. The method of claim 1, wherein each pattern further comprises one or more calls placed and a time stamp at which each of the one or more calls placed is performed.

3. The method of claim 2, wherein the time stamp at which each web access is performed is within a period of time from the time stamp at which each of the one or more calls placed is performed.

4. The method of claim 3, wherein the identifying one or more patterns among the web log records for the plurality of call log records comprises combining patterns into groups of patterns based upon the period of time from the time stamp at which each of the one or more calls placed is performed and the time stamp at which each of the one or more web accesses is performed.

5. The method of claim 1, wherein each web log record further comprises a plurality of clicks arranged in a sequence, and wherein each pattern further comprises a sequence of clicks for one or more web accesses.

6. The method of claim 5, wherein the identifying one or more patterns among the web log records for the plurality of call log records comprises combining patterns into groups of patterns based upon patterns having a similar sequence of clicks.

7. A system for performing data mining, the system comprising:

a processing device;

a non-transitory computer-readable storage medium in communication with the processing device,

wherein the computer-readable storage medium contains one or more programming instructions that, when executed, cause the processing device to;

for each of a plurality of call log records, associate one or more web log records with the call log record, wherein each web log record comprises a web access and a time stamp, wherein each call log record comprises a call topic and a time stamp, and wherein each of the one or more web log records and the call log record are associated with a user,

identify one or more patterns among the web log records for the plurality of call log records, wherein each pattern comprises one or more web accesses, a time stamp at which each web access is performed and the call topic for the call log record,

identify one or more web log records associated with a new call, and

predict a call topic for the new call based on at least one pattern and the one or more web log records.

8. The system of claim 7, wherein each pattern further comprises one or more calls placed and a time stamp at which each of the one or more calls placed is performed.

9. The system of claim 8, wherein the time stamp at which each web access is performed is within a period of time from the time stamp at which each of the one or more calls placed is performed.

10. The system of claim 9, wherein the identifying one or more patterns among the web log records for the plurality of call log records comprises combining patterns into groups of patterns based upon the period of time from the time stamp at which each of the one or more calls placed is performed and the time stamp at which each of the one or more web accesses is performed.

11. The system of claim 7, wherein each web log record further comprises a plurality of clicks arranged in a sequence, and wherein each pattern further comprises a sequence of clicks for one or more web accesses.

12. The system of claim 11, wherein the computer-readable storage medium contains one or more programming instructions that, when executed, cause the processing device to identify one or more patterns among the web log records for the plurality of call log records, and further combine patterns into groups of patterns based upon patterns having a similar sequence of clicks.

13. A method of predicting a topic of a call, the method comprising:

providing a web log having a plurality of web log records, wherein each web log record comprises a web access, a time stamp and an associated user;

providing a call log having a plurality of call log records, wherein each call log record comprises a call topic, a time stamp and an associated user;

for each call log record, associating, by a processing device, one or more web log records with the call log record, wherein the associated user for each of the associated one or more web log records and the call log record are the same;

identifying, by the processing device, one or more patterns among the web log records for the plurality of call log records, wherein each pattern comprises one or more web accesses, a time stamp at which each of the one or more web accesses is performed and the call topic for the call log record;

identifying, by the processing device, one or more web log records associated with a new call; and

predicting, by the processing device, a call topic for the new call based on at least one pattern and the one or more web log records.

14. The method of claim 13, wherein each pattern further comprises one or more calls placed and a time stamp at which each of the one or more calls placed is performed.

15. The method of claim 14, wherein the time stamp at which each web access is performed is within a period of time from the time stamp at which each of the one or more calls placed is performed.

16. The method of claim 15, wherein the identifying one or more patterns among the web log records for the plurality of call log records comprises combining patterns into groups of patterns based upon the period of time from the time stamp at which each of the one or more calls placed is performed and the time stamp at which each of the one or more web accesses is performed.

17. The method of claim 13, wherein each web log record further comprises a plurality of clicks arranged in a sequence, and wherein each pattern further comprises a sequence of clicks for one or more web accesses.

18. The method of claim 17, wherein the identifying one or more patterns among the web log records for the plurality of call log records comprises combining patterns into groups of patterns based upon patterns having a similar sequence of clicks.