Systems and methods for determining an article association measure

• US 8,667,013 B1
• Filed: 01/27/2012
• Issued: 03/04/2014
• Est. Priority Date: 06/30/2004
• Status: Expired due to Fees

• Current Assignee: Google Inc. (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Lawrence, Stephen R., Wang, Niniane
• Primary Examiner(s): Alam, Hosain
• Assistant Examiner(s): Ahluwalia, Navneet K
• Application Number:
  US13/360,593
• Time in Patent Office: 767 Days
• Field of Search: 707770-775, 707/758
• US Class Current: 707/771
• CPC Class Codes: G06F 16/903 Querying for retrieval from the web G06F16/953
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques

Abstract

Systems and methods for determining an article association measure are described. In one described method, a computer program receives a first article identifier associated with a first attribute and a second article identifier associated with a second attribute and determines a similarity measure of the first article identifier to the second article identifier based at least in part on the first attribute and on the second attribute. The similarity measure may provide a means for associating various article identifiers so that they may be, for example, retrieved and/or ranked in subsequent search queries.

Claims

1. A computer-implemented method comprising:

accessing data describing a plurality of captured events, wherein the events comprise interactions with articles;

for a first article identifier identifying a first article and associated with a first date-time attribute of a first one of the plurality of events, and for a second article identifier identifying a second article and associated with a second date-time attribute of a second one of the plurality of events;

determining an association measure of the first article identifier to the second article identifier based at least in part on the first date-time attribute and on the second date-time attribute, the determining comprising;

determining a difference between the first date-time attribute associated with the first article identifier and the second date-time attribute associated with the second article identifier,

comparing the difference between the first date-time attribute associated with the first article identifier and the second date-time attribute associated with the second article identifier to a threshold, and

setting the association measure of the first article identifier to the second article identifier based at least in part on whether the difference between the first date-time attribute associated with the first article identifier and the second date-time attribute associated with the second article identifier exceeds the threshold;

receiving a query;

receiving search results as a result of query execution, the search results comprising the first article identifier and omitting the second article identifier;

identifying the second article identifier as being related to the first article identifier based at least in part on the association measure; and

responsive to identifying the second article identifier as being related to the first article identifier, including the second article identifier as part of the received search results.

2. The method of claim 1, wherein the difference varies inversely with the association measure.

3. The method of claim 1, wherein the association measure comprises a plurality of association measures, and further comprising combining the plurality of association measures to create a combined association measure.

4. The method of claim 3, wherein combining the plurality of association measures to create the combined association measure comprises weighting each association measure of the plurality of association measures according to a weight corresponding to a type of the association measure.

5. The method of claim 1, further comprising normalizing the association measure to create a normalized association measure.

6. The method of claim 1, wherein the association measure is computed using one of a linear, piecewise, or exponential function, wherein the measure indicates higher association for smaller differences.

7. The method of claim 1, wherein the first article identifier is further associated with a first content attribute and the second article identifier is further associated with a second content attribute, the method further comprising determining the association measure of the first article identifier to the second article identifier based at least in part on the first content attribute and on the second content attribute.

8. The method of claim 1, wherein the first article identifier is further associated with a first metadata attribute and the second article identifier is further associated with a second metadata attribute, the method further comprising determining the association measure of the first article identifier to the second article identifier based at least in part on the first metadata attribute and on the second metadata attribute.

9. The method of claim 8, wherein the first metadata attribute comprises an author.

10. The method of claim 8, wherein the first metadata attribute comprises an organization.

11. The method of claim 1, wherein the association score is determined by an indexer as part of indexing and applied as part of execution of a query.

12. The method of claim 1, wherein the association score is determined as part of execution of a query.

13. A non-transitory computer-readable storage medium on which is encoded executable program code, the program code comprising:

program code for accessing data describing a plurality of captured events, wherein the events comprise interactions with articles;

program code for, for a first article identifier identifying a first article and associated with a first date-time attribute of a first one of the plurality of events, and for a second article identifier identifying a second article and associated with a second date-time attribute of a second one of the plurality of events;

determining an association measure of the first article identifier to the second article identifier based at least in part on the first date-time attribute and on the second date-time attribute, the determining comprising;

determining a difference between the first date-time attribute associated with the first article identifier and the second date-time attribute associated with the second article identifier,

comparing the difference between the first date-time attribute associated with the first article identifier and the second date-time attribute associated with the second article identifier to a threshold, and

setting the association measure of the first article identifier to the second article identifier based at least in part on whether the difference between the first date-time attribute associated with the first article identifier and the second date-time attribute associated with the second article identifier exceeds the threshold;

program code for receiving a query;

program code for receiving search results as a result of query execution, the search results comprising the first article identifier and omitting the second article identifier;

program code for identifying the second article identifier as being related to the first article identifier based at least in part on the association measure; and

program code for, responsive to identifying the second article identifier as being related to the first article identifier, including the second article identifier as part of the received search results.

14. The computer-readable storage medium of claim 13, wherein the association measure comprises a plurality of association measures, and further comprising program code for combining the plurality of association measures to create a combined association measure.

15. The computer-readable storage medium of claim 13, further comprising program code for normalizing the association measure to create a normalized association measure.

16. A computer system, comprising:

a computer processor; and

a non-transitory computer-readable medium storing processor-executable program instructions that when executed by the processor perform actions comprising;

accessing data describing a plurality of captured events, wherein the events comprise interactions with articles;

for a first article identifier identifying a first article and associated with a first date-time attribute of a first one of the plurality of events, and for a second article identifier identifying a second article and associated with a second date-time attribute of a second one of the plurality of events;

determining an association measure of the first article identifier to the second article identifier based at least in part on the first date-time attribute and on the second date-time attribute, the determining comprising;

determining a difference between the first date-time attribute associated with the first article identifier and the second date-time attribute associated with the second article identifier,

comparing the difference between the first date-time attribute associated with the first article identifier and the second date-time attribute associated with the second article identifier to a threshold, and

setting the association measure of the first article identifier to the second article identifier based at least in part on whether the difference between the first date-time attribute associated with the first article identifier and the second date-time attribute associated with the second article identifier exceeds the threshold;

receiving a query;

receiving search results as a result of query execution, the search results comprising the first article identifier and omitting the second article identifier;

identifying the second article identifier as being related to the first article identifier based at least in part on the association measure; and

responsive to identifying the second article identifier as being related to the first article identifier, including the second article identifier as part of the received search results.

17. The computer system of claim 16, wherein the first article identifier is further associated with a first content attribute and the second article identifier is further associated with a second content attribute, the actions further comprising determining the association measure of the first article identifier to the second article identifier based at least in part on the first content attribute and on the second content attribute.

18. The computer system of claim 16, wherein the first article identifier is further associated with a first metadata attribute and the second article identifier is further associated with a second metadata attribute, the actions further comprising determining the association measure of the first article identifier to the second article identifier based at least in part on the first metadata attribute and on the second metadata attribute.

19. The computer system of claim 18, wherein the first metadata attribute comprises an author.