Use of web usage trail data to identify relationships between browsable items

• US 7,159,023 B2
• Filed: 12/16/2003
• Issued: 01/02/2007
• Est. Priority Date: 12/23/1999
• Status: Expired due to Fees

• Current Assignee: Alexa Internet Inc (Amazon.com, Inc.)
• Original Assignee: Alexa Internet Inc (Amazon.com, Inc.)
• Inventors: Tufts, Patrick
• Primary Examiner(s): Lim; Krisna
• Application Number:
  US10/737,168
• Publication Number:
  US 20040205516A1
• Time in Patent Office: 1,113 Days
• Field of Search: 709/224, 709/223, 705/52
• US Class Current: 709/224
• CPC Class Codes: H04L 67/53 using third party service providers
• CPC Class Codes: H04L 69/329 in the application layer [OSI layer 7]
• CPC Class Codes: H04L 9/40 Network security protocols

Abstract

Related web locations are identified from the web usage trails of a population of users. Each usage trail is preferably in the general form of a sequence of URLs accessed by a user during a browsing session. The usage trails are preferably collected using a special client application that operates in conjunction with a web browser and reports the navigation activities of users. In a preferred embodiment, the relatedness of two web locations A and B is determined using a minimum sensitivity calculation that takes into consideration the number of transitions that occurred between A and B relative to the total number of transitions that involved A and/or B. The method may be used independently, or may be used in combination with other methods for identifying related web locations.

Claims

1. A computer system, comprising:

a usage trail data accumulator configured to collect usage trail data descriptive of usage trails of a plurality of users, said usage trail data identifying sequences of network resources accessed by corresponding users during network browsing sessions of such users; and

a metadata generator configured to analyze the usage trail data collected by the usage trail data accumulator and to generate scores that represent degrees of relationship between specific network resources, said scores being normalized values falling within a bounded range, and being dependent upon the usage trails of the plurality of users.

2. The computer system of claim 1, wherein the metadata generator is configured to generate a score that represents a degree of relationship between a first network resource and a second network resource in-part by determining a total number of transitions that occurred between the first network resource and the second network resource, as determined from the usage trail data.

3. The computer system of claim 2, wherein the first and second network resources are respective web sites.

4. The computer system of claim 2, wherein the first and second network resources are respective web pages.

5. The computer system of claim 2, wherein the metadata generator is configured to determine said total number of transitions such that at least some multi-step transitions are counted.

6. The computer system of claim 2, wherein the metadata generator is configured to determine said total number of transitions such that only N-step to M-step transitions are counted, where N and M are selected integers.

7. The computer system of claim 2, wherein the score is further dependent on a total number of transitions that occurred in said usage trails between the first network resource and all other network resources of a like type.

8. The computer system of claim 7, wherein the score is further dependent on a total number of transitions that occurred in said usage trails between the second network resource and all other network resources of a like type.

9. The computer system of claim 2, wherein the score is further dependent upon whether a direct link exists between the first and second network resources.

10. The computer system of claim 2, wherein the metadata generator is configured to generate said score taking into consideration, for a given usage trail that includes an access to the first network resource and an access to the second network resource, an amount of time that elapsed between the access to the first network resource and the access to the second network resource.

11. The computer system of claim 1, wherein the metadata generator uses a minimum sensitivity calculation to generate the scores.

12. The computer system of claim 1, wherein the usage trail data accumulator is configured to collect the usage trail data using a browser plug-in that runs on user computers and reports browsing actions of corresponding users.

13. The computer system of claim 1, wherein the bounded range is 0 to 1.

14. A computer-implemented data mining method, comprising:

storing usage trail data descriptive of usage trails of a plurality of users, each usage trail identifying a sequence of network resources accessed by a corresponding user during a browsing session; and

programmatically generating a score that represents a degree to which a first network resource and a second network resource are related to each other, said score falling within a predefined, bounded range, and being dependent upon a total number of transitions in said usage trails between the first network resource and the second network resource as determined from the usage trail data, whereby the score reflects browsing behaviors of users.

15. The method of claim 14, wherein the first and second network resources are respective web sites.

16. The method of claim 14, wherein the first and second network resources are respective web pages.

17. The method of claim 14, wherein said total number of transitions is determined such that at least some multi-step transitions are counted.

18. The method of claim 14, wherein the total number of transitions is determined such that only N-step to M-step transitions are counted, where N and M are selected integers.

19. The method of claim 14, wherein the score is further dependent on a total number of transitions in said usage trails between the first network resource and all other network resources of a like type.

20. The method of claim 19, wherein the score is further dependent on a total number of transitions that occurred in said usage trails between the second network resource and all other network resources of a like type.

21. The method of claim 14, wherein programmatically generating the score comprises using a minimum sensitivity calculation to generate the score.

22. The method of claim 14, wherein programmatically generating the score comprises taking into consideration whether a direct link exists between the first and second network resources, as determined from an automated analysis of web page content.

23. The method of claim 14, wherein programmatically generating the score comprises taking into consideration, for a given usage trail that includes an access to the first network resource and an access to the second network resource, an amount of time that elapsed between the access to the first network resource and the access to the second network resource.

24. The method of claim 14, further comprising collecting the usage trail data using a browser plug-in that runs on user computers and reports browsing actions of said users.

25. The method of claim 14, wherein the predefined, bounded range is 0 to 1.