UserRank: ranking linked nodes leveraging user logs

• US 7,693,818 B2
• Filed: 11/15/2005
• Issued: 04/06/2010
• Est. Priority Date: 11/15/2005
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Majumder, Rangan
• Primary Examiner(s): Rones; Charles
• Assistant Examiner(s): Quader; Fazlul
• Application Number:
  US11/274,699
• Publication Number:
  US 20070112768A1
• Time in Patent Office: 1,603 Days
• Field of Search: 707/2
• US Class Current: 1/1
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: Y10S 707/99932 Access augmentation or optimizing

Abstract

The claimed subject matter provides a system and/or a method that facilitates utilizing transition probability in static rankings associated with at least one document. An interface can receive data related to a query, wherein the query can be associated with a search from a user. A rank component can provide query results that are prioritized utilizing a transition probability based on user activity included within a user log.

Claims

1. A system that facilitates providing query results to a user, comprising:

a processor;

system memory;

an interface that receives data related to a query associated with a specific user;

a user log component that comprises at least a repository of historic activities, behaviors, or combinations thereof specific to a user, wherein the logged historic user activities and/or behaviors are related to prospective interactions with objects related to the query results;

a rank component that provides ranked query results adapted in relation to a specific user log, wherein the query results are at least in part prioritized utilizing a transition probability determination related to the specific user for user transitions between objects related to prospective query results based on historical user activity, user behavior, or combinations thereof, wherein the transition probability determination comprises utilizing a weighting technique that implements the following representation of a Markov model;

2. The system of claim 1, the transition probability determination is related to the specific user for user transitions between documents related to prospective query results based on historical user activity, user behavior, or combinations thereof.

3. The system of claim 2, the document is at least one of the following:

a hypertext markup language (HTML) document;

a web page;

a link;

a hyperlink;

a uniform resource locator (URL); and

a resource on the World Wide Web.

4. The system of claim 1, the user activity is-at least one of the following:

a document request;

a document referrer; and

a document request uniform resource identifier (URI).

5. The system of claim 1, the user activity is at least one of time on and object, entering an object, exiting an object, logging into an object, logging out of an object, linking to a disparate object from an object, a log off associated with a web page, or combinations thereof.

6. The system of claim 1, further comprising a selector component that analyzes available data within the user log to identify optimal data to indicate user activity.

7. The system of claim 6, further comprising a filter component that sifts data based at least in part upon the optimal data identified.

8. The system of claim 7, the filter component utilizes at least one of a hash technique and a sort technique to group the data to provide aggregate counts related to the object.

9. The system of claim 1, further comprising a weight component that implements accruing weight to the query result based at least in part upon the transition probability.

10. The system of claim 9, the weight component utilizes a first order Markov chain that is trained on a sequence of transitions that occurred between at least two documents linked via a hyperlink.

11. The system of claim 1, further comprising a browser that allows a user to access at least one object and updates the user log with each access.

12. The system of claim 1, the user log is an Internet information server (IIS) user log.

13. A computer-implemented method that facilitates prioritizing query results, comprising:

receiving a user log including data related to historic user activities, behaviors, or combinations thereof for historic user interaction with documents, said documents being related to the prospective interaction with documents related to the query results;

utilizing a weighting technique to identify a transition probability related to the specific user for prospective interactions with document related to query results based at least in part on a user activity and/or behavior associated with the user log, wherein the weighting technique utilizes the following representation of a Markov model;

14. The method of claim 13, the weighting technique is a first order Markov model of a web graph from the user log.

15. The method of claim 13, the weighting technique computes a static rank of each document utilizing the transition probability form each node to a neighboring node.

16. A computer-implemented system that facilitates query results to a user, comprising:

a processing means;

a data storage means;

means for receiving data related to a query;

means for logging historical user data related to user actions, behaviors, or combinations thereof to facilitate probabilistic determinations for future user behaviors and/or actions related to the objects of the same or similar queries;

means for utilizing a utilizing a weighting technique to identify a transition probability related to the specific user for prospective interactions with document related to query results based at least in part on a user activity and/or behavior associated with the user log, wherein the weighting technique utilizes the following representation of a Markov model;