Sampling internet user traffic to improve search results

• US 8,046,357 B2
• Filed: 11/07/2005
• Issued: 10/25/2011
• Est. Priority Date: 11/07/2005
• Status: Expired due to Fees

• Current Assignee: IAC Search & Media Incorporated (IAC/InterActiveCorp)
• Original Assignee: IAC Search & Media Incorporated (IAC/InterActiveCorp)
• Inventors: Mori, Monica, Gulli, Antonino, Savona, Antonio
• Primary Examiner(s): Timblin; Robert
• Assistant Examiner(s): Khakhar; Nirav K
• Application Number:
  US11/268,996
• Publication Number:
  US 20070112730A1
• Time in Patent Office: 2,178 Days
• Field of Search: 707/726, 707/999.007
• US Class Current: 707/726
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques

Abstract

Methods, systems and apparatus for improving Internet search results include monitoring network activity on a communications network, generating a near real-time map of the network activity, and integrating the near real-time map with a search engine.

Claims

1. A method, comprising:

monitoring network activity in an activity log at a network access node on a communications network, wherein the communications network comprises the World Wide Web (WWW), and wherein the network access node comprises one of a web proxy server and an Internet service provider (ISP), and wherein the network activity comprises web traffic on the WWW;

transferring network activity data from the activity log to a network traffic monitor via a proprietary data link;

generating a near real-time map of the network activity from the network activity data, wherein generating the near real-time map of the network activity comprises generating a web graph from the web traffic, the web graph comprising a plurality of web nodes linked by a plurality of web edges, wherein each of the plurality of web nodes represents a corresponding web object and each of the plurality of web edges represents a weighted link from a first web object to a second web object, wherein a weight of the weighted link is based on a frequency of navigation between the first web object and the second web object;

integrating the near real-time map of the network activity with search engine data acquired via a nonproprietary data link;

receiving a search request from a user;

extracting search results in response to the search request according to their relevance to the search request from the search engine data including the integration of the near real-time map of the network activity and the search engine data;

preparing a list of search results wherein the highest ranked search results are placed at the beginning of the list, the ranking being based on the respective weights of the web edges; and

returning the list of ranked search results to the user.

2. The method of claim 1, wherein integrating the near real-time map of the network activity with the search engine data comprises:

updating a keyword index of the search engine data with the network activity data;

adding and deleting web objects in the search engine data based on the web graph; and

clustering web objects and search engine queries based on the web graph.

3. The method of claim 1, further comprising:

creating a web edge when a user navigates between the first web object and the second web object without a pre-existing web edge; and

extinguishing the web edge after a specified period of disuse.

4. The method of claim 1, further comprising:

incrementing a weight of the web edge each time a hyperlink in the first web object is used to navigate from the first web object to the second web object; and

decrementing the weight of the web edge after the specified period of disuse.

5. The method of claim 1, further comprising:

incrementing a weight of the web edge each time the user navigates from the first web object to the second web object without using a hyperlink; and

decrementing the weight of the web edge after the specified period of disuse.

6. The method of claim 4, further comprising anchoring the web edge with an anchor comprising a portion of text surrounding the hyperlink in the first web object.

7. The method of claim 6, wherein a search request invokes at least some portion of the anchor, further comprising ranking a search result in a search engine based on the weight of the web edge linking the first web object to the second web object.

8. The method of claim 1, wherein monitoring-the network activity comprises extracting HTTP_REFER tags from the activity log at the network access node.

9. The method of claim 1, wherein a web object in the web graph is connected to one or more web edges, wherein a weight of the web object corresponds to the weights of the one or more web edges, further comprising adding the web object to a search engine index when the weight of the web object exceeds a predetermined threshold, and deleting the web object from the search engine index when the weight of the web object falls below the predetermined threshold.

10. The method of claim 9, further comprising generating a relevance score for the web object based on the weight of the web object and a ranking of the web object in a search result.

11. The method of claim 9, further comprising deleting the web object from the search engine index when the web traffic includes missing object errors for the web object.

12. The method of claim 1, wherein the web traffic includes a query and a pick in a first search engine, wherein the pick corresponds to a web object, the method further comprising associating the web object with the query in a second search engine.

13. The method of claim 1, wherein the web traffic includes a first query and a first pick in a first search engine corresponding to a web object and a second query and a second pick in a second search engine corresponding to the web object, the method further comprising clustering the first query and the second query in at least the first search engine.

14. The method of claim 1, wherein the web traffic comprises a burst of queries directed toward a set of web objects, the method further comprising clustering the set of web objects and the burst of queries in the search engine while the burst of queries is sustained above a predetermined threshold.

15. The method of claim 1, wherein the web graph is associated with the web traffic of a definable group of users having one or more web traffic metrics in common during a monitoring period, the method further comprising biasing the search results of a new user with the web graph when the new user exhibits web activity that is similar to the web traffic metrics of the definable group of users.

16. The method of claim 15, wherein the biasing of the search results has a defined lifetime related to the monitoring period.

17. A system, comprising:

a memory;

a communications port coupled to a network access node via a proprietary data link and to a communications network via a non-proprietary data link; and

a processing device coupled with the memory and the communications port, the processing device configured as a search engine and further configured to;

monitor network activity data in an activity log at an access node on a communications network, wherein the communications network comprises the World Wide Web (WWW), and wherein the network access node comprises one of a web proxy server and an Internet service provider (ISP), and wherein the network activity comprises web traffic on the WWW,

transfer network activity data from the network access node to the memory via the proprietary data link;

generate a near real-time map of the network activity from the network activity data, wherein to generate a near real-time map of the network activity the processing device is configured to generate a web graph from the web traffic, the web graph comprising a plurality of web nodes linked by a plurality of web edges, wherein each of the plurality of web nodes represents a corresponding web object and each of the plurality of web edges represents a weighted link from a first web object to a second web object, wherein a weight of the weighted link is based on a frequency of navigation between the first web object and the second web object ;

integrate the near real-time map of the network activity with search engine data acquired via the nonproprietary data link;

receive a search request from a user;

extract the search results in response to the search request according to their relevance to the search request from the search engine data including the integration of the near real-time map of the network activity and the search engine data;

prepare a list of search results wherein the highest ranked search results are placed at the beginning of the list, the ranking being based on the respective weights of the web edges; and

return the list of ranked search results to the user.

18. The system of claim 17, wherein to integrate the near real-time map of the network activity with the search engine data, the processing device is configured to:

update a keyword index of the search engine data with the network activity data;

add and delete web objects in the search engine data based on the web graph; and

cluster web objects and search engine queries based on the web graph.

19. The system of claim 17, wherein the processing device is further configured to:

create a web edge when a user navigates between the first web object and the second web object without a pre-existing web edge; and

extinguish the web edge after a specified period of disuse.

20. The system of claim 19, wherein the processing device is further configured to anchor the web edge with an anchor comprising a portion of text surrounding the hyperlink in the first web object.

21. The system of claim 20 wherein a search request invokes at least some portion of the anchor, wherein the processing device is further configured to rank a search result in a search engine based on the weight of the web edge linking the first web object to the second web object.

22. The system of claim 17, wherein the processing device is further configured to:

increment a weight of the web edge each time a hyperlink in the first web object is used to navigate from the first web object to the second web object; and

decrement the weight of the web edge after the specified period of disuse.

23. The system of claim 17, wherein the processing device is further configures to:

increment a weight of the web edge each time the user navigates from the first web object to the second web object without using a hyperlink; and

decrement the weight of the web edge after the specified period of disuse.

24. The system of claim 17, wherein to record the network activity, the processing device is configured to extract HTTP_REFER tags from the activity log at the network access node.

25. The system of claim 24, wherein the processing device is further configured to generate a relevance score for the web object based on a weight of the web object and a ranking of the web object in a search result.

26. The system of claim 24, wherein the processing device is further configured to delete the web object from the search engine index when the web traffic includes missing object errors for the web object.

27. The system of claim 17, wherein a web object in the web graph is connected to one or more web edges, wherein a weight of the web object corresponds to the weights of the one or more web edges, wherein the processing device is further configured to add the web object to a search engine index when the weight of the web object exceeds a predetermined threshold, and delete the web object from the search engine index when the weight of the web object falls below the predetermined threshold.

28. The system of claim 17, wherein the web traffic includes a query and a pick in a first search engine, wherein the pick corresponds to a web object, wherein the processing device is further configured to associate the web object with the query in a second search engine.

29. The system of claim 17, wherein the web traffic includes a first query and a first pick in a first search engine corresponding to a web object and a second query and a second pick in a second search engine corresponding to the web object, wherein the processing device is further configured to cluster the first query and the second query in at least the first search engine.

30. The system of claim 17, wherein the web traffic comprises a burst of queries directed toward a set of web objects, wherein the processing device is further configured to cluster the set of web objects and the burst of queries in the search engine while the burst of queries is sustained above a predetermined threshold.

31. The system of claim 17, wherein the web graph is associated with the web traffic of a definable group of users having one or more web traffic metrics in common during a monitoring period, wherein the processing device is further configured to bias the search results of a new user with the web graph when the new user exhibits web activity that is similar to the web traffic metrics of the definable group of users.

32. The system of claim 31, wherein the bias of the search results has a defined lifetime related to the monitoring period.

33. An article of manufacture, comprising a machine-readable storage medium including machine-accessible instructions that when accessed by a data processing system, cause the data processing system to perform operations, comprising:

monitoring network activity in an activity log at a network access node on a communications network, wherein the communications network comprises the World Wide Web (WWW), and wherein the network access node comprises one of a web proxy server and an Internet service provider (ISP), and wherein the network activity comprises web traffic on the WWW;

transferring network activity data from the activity log to a network traffic monitor via a proprietary data link;

generating a near real-time map of the network activity from the network activity data, wherein generating the near real-time map of the network activity comprises generating a web graph from the web traffic, the web graph comprising a plurality of web nodes linked by a plurality of web edges, wherein each of the plurality of web nodes represents a corresponding web object and each of the plurality of web edges represents a weighted link from a first web object to a second web object, wherein a weight of the weighted link is based on a frequency of navigation between the first web object and the second web object;

integrating the near real-time map of the network activity with search engine data acquired via a nonproprietary data link;

receiving a search request from a user;

extracting search results in response to the search request according to their relevance to the search request from the search engine data including the integration of the near real-time map of the network activity and the search engine data;

preparing a list of search results wherein the highest ranked search results are placed at the beginning of the list, the ranking being based on the respective weights of the web edges; and

returning the list of ranked search results to the user.

34. The article of manufacture of claim 33, wherein integrating the near real-time map of the network activity with the search engine data comprises:

updating a keyword index of the search engine data with the network activity data;

adding and deleting web objects in the search engine data based on the web graph; and

clustering web objects and search engine queries based on the web graph.

35. The article of manufacture of claim 33, further including instructions that, when accessed by the data processing system, cause the data processing system to perform operations, comprising:

creating a web edge when a user navigates between the first web object and the second web object without a pre-existing web edge; and

extinguishing the web edge after a specified period of disuse.

36. The article of manufacture of claim 33, further including instructions that, when accessed by the data processing system, cause the data processing system to perform operations, comprising:

incrementing a weight of the web edge each time a hyperlink in the first web object is used to navigate from the first web object to the second web object; and

decrementing the weight of the web edge after the specified period of disuse.

37. The article of manufacture of claim 36, further including instructions that, when accessed by the data processing system, cause the data processing system to perform operations, comprising:

anchoring the web edge with an anchor comprising a portion of text surrounding the hyperlink in the first web object.

38. The article of manufacture of claim 37, wherein a search request invokes at least some portion of the anchor, further including instructions that, when accessed by the data processing system, cause the data processing system to perform operations, comprising:

ranking a search result in a search engine based on the weight of the web edge linking the first web object to the second web object.

39. The article of manufacture of claim 33, further including instructions that, when accessed by the data processing system, cause the data processing system to perform operations, comprising:

incrementing a weight of the web edge each time the user navigates from the first web object to the second web object without using a hyperlink; and

decrementing the weight of the web edge after the specified period of disuse.

40. The article of manufacture of claim 33, wherein recording the network activity comprises extracting HTTP_REFER tags from the activity log at the network access node.

41. The article of manufacture of claim 33, wherein a web object in the web graph is connected to one or more web edges, wherein a weight of the web object corresponds to the weights of the one or more web edges, further including instructions that, when accessed by the data processing system, cause the data processing system to perform operations, comprising:

adding the web object to a search engine index when the weight of the web object exceeds a predetermined threshold; and

deleting the web object from the search engine index when the weight of the web object falls below the predetermined threshold.

42. The article of manufacture of claim 41, further including instructions that, when accessed by the data processing system, cause the data processing system to perform operations, comprising:

generating a relevance score for the web object based on the weight of the web object and a ranking of the web object in a search result.

43. The article of manufacture of claim 41, further including instructions that, when accessed by the data processing system, cause the data processing system to perform operations, comprising:

deleting the web object from the search engine index when the web traffic includes missing object errors for the web object.

44. The article of manufacture of claim 33, wherein the web traffic includes a query and a pick in a first search engine, wherein the pick corresponds to a web object, further including instructions that, when accessed by the data processing system, cause the data processing system to perform operations, comprising:

associating the web object with the query in a second search engine.

45. The article of manufacture of claim 33, wherein the web traffic includes a first query and a first pick in a first search engine corresponding to a web object and a second query and a second pick in a second search engine corresponding to the web object, further including instructions that, when accessed by the data processing system, cause the data processing system to perform operations, comprising:

clustering the first query and the second query in at least the first search engine.

46. The article of manufacture of claim 33, wherein the web traffic comprises a burst of queries directed toward a set of web objects, further including instructions that, when accessed by the data processing system, cause the data processing system to perform operations, comprising:

clustering the set of web objects and the burst of queries in the search engine while the burst of queries is sustained above a predetermined threshold.

47. The article of manufacture of claim 33, wherein the web graph is associated with the web traffic of a definable group of users having one or more web traffic metrics in common during a monitoring period, further including instructions that, when accessed by the data processing system, cause the data processing system to perform operations, comprising:

biasing the search results of a new user with the web graph when the new user exhibits web activity that is similar to the web traffic metrics of the definable group of users.

48. The article of manufacture of claim 47, wherein the biasing of the search results has a defined lifetime related to the monitoring period.