Methods and systems for improving a search ranking using related queries

• US 8,380,705 B2
• Filed: 08/18/2011
• Issued: 02/19/2013
• Est. Priority Date: 09/12/2003
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Tong, Simon, Pearson, Mark, Brin, Sergey
• Primary Examiner(s): PARDO, THUY N
• Application Number:
  US13/212,726
• Publication Number:
  US 20120011117A1
• Time in Patent Office: 551 Days
• Field of Search: 707/721, 707/748, 707/749, 707/751, 707/765
• US Class Current: 707/721
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 16/9538 Presentation of query results
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access
• CPC Class Codes: Y10S 707/99942 Manipulating data structure, e.g. compression, compaction, compilation
• CPC Class Codes: Y10S 707/99943 Generating database or data structure, e.g. via user interface
• CPC Class Codes: Y10S 707/99944 Object-oriented database structure
• CPC Class Codes: Y10S 707/99945 Object-oriented database structure processing

Abstract

Systems and methods that improve search rankings for a search query by using data associated with queries related to the search query are described. In one aspect, a search query is received, a related query related to the search query is determined, an article (such as a web page) associated with the search query is determined, and a ranking score for the article based at least in part on data associated with the related query is determined. Several algorithms and types of data associated with related queries useful in carrying out such systems and methods are described.

Claims

1. A computer-implemented method comprising:

determining, by a computer system, a score for a document as a result for a search query, the score based, at least in part, on one or more related search queries that have at least a threshold relationship to the search query, wherein determining the score comprises;

identifying click data associated with the document and the one or more related search queries, wherein the click data indicates how frequently the document was selected when the document was presented in search results for the one or more related search queries over a period of time;

weighting the click data based on weights for the one or more related search queries, wherein the weights indicate how strongly the one or more related queries are related to the search query; and

combining the weighted click data to generate at least a component of the score; and

providing, by the computer system, the score for the document as a result for the search query.

2. The method of claim 1, further comprising, for a particular related search query from the one or more related search queries, normalizing the weighted click data for the document and the particular related search query by dividing the weighted click data for the document and the particular related search query by a number of times the related search query was received over the period of time;

wherein combining the weighted click data comprises combining normalized weighted click data for the document and the one or more related search queries.

3. The method of claim 2, wherein combining the normalized weighted click data comprises adding together the normalized weighted click data for the document and the one or more related search queries.

4. The method of claim 1, further comprising combining additional click data with the combined weighted click data to generate the score, wherein the additional click data indicates how frequently the document was selected when the document was presented in search results for the search query.

5. The method of claim 4, wherein the combined weighted click data is adjusted by a blending factor before it is combined with the additional click data, wherein the blending factor indicates a degree to which the click data for the document and the one or more related search queries reliably indicates user behavior with regard to the search query.

6. The method of claim 5, wherein combining the additional click data with the combined weighted click data comprises adding i) the combined weighted click data multiplied by the blending factor and ii) the additional click data multiplied by one minus the blending factor.

7. The method of claim 1, wherein the click data corresponds to a number of times that the document was selected when it was presented in the search results for the one or more related search queries over the period of time.

8. A computer program product tangibly embodied on a computer readable medium storing instructions that, when executed by one or more computing devices, cause operation to be performed comprising:

determining a score for a document as a result for a search query, the score based, at least in part, on one or more related search queries that have at least a threshold relationship to the search query, wherein determining the score comprises;

identifying click data associated with the document and the one or more related search queries, wherein the click data indicates how frequently the document was selected when the document was presented in search results for the one or more related search queries over a period of time;

weighting the click data based on weights for the one or more related search queries, wherein the weights indicate how strongly the one or more related queries are related to the search query; and

combining the weighted click data to generate at least a component of the score; and

providing the score for the document as a result for the search query.

9. The computer program product of claim 8, wherein the operations further comprise, for a particular related search query from the one or more related search queries, normalizing the weighted click data for the document and the particular related search query by dividing the weighted click data for the document and the particular related search query by a number of times the related search query was received over the period of time;

wherein combining the weighted click data comprises combining normalized weighted click data for the document and the one or more related search queries.

10. The computer program product of claim 9, wherein combining the normalized weighted click data comprises adding together the normalized weighted click data for the document and the one or more related search queries.

11. The computer program product of claim 8, wherein the operations further comprise combining additional click data with the combined weighted click data to generate the score, wherein the additional click data indicates how frequently the document was selected when the document was presented in search results for the search query.

12. The computer program product of claim 11, wherein the combined weighted click data is adjusted by a blending factor before it is combined with the additional click data, wherein the blending factor indicates a degree to which the click data for the document and the one or more related search queries reliably indicates user behavior with regard to the search query.

13. The computer program product of claim 12, wherein combining the additional click data with the combined weighted click data comprises adding i) the combined weighted click data multiplied by the blending factor and ii) the additional click data multiplied by one minus the blending factor.

14. The computer program product of claim 8, wherein the click data corresponds to a number of times that the document was selected when it was presented in the search results for the one or more related search queries over the period of time.

15. A system comprising:

a computer system; and

a related query processor of the computer system to determine a score for a document as a result for a search query, the score based, at least in part, on one or more related search queries that have at least a threshold relationship to the search query, wherein the score is determined by performing operations comprising;

identifying click data associated with the document and the one or more related search queries, wherein the click data indicates how frequently the document was selected when the document was presented in search results for the one or more related search queries over a period of time;

weighting the click data based on weights for the one or more related search queries, wherein the weights indicate how strongly the one or more related queries are related to the search query; and

combining the weighted click data to generate at least a component of the score;

wherein the related query processor is further configured to provide the score for the document as a result for the search query to a ranking processor.

16. The system of claim 15, wherein the operations further comprise, for a particular related search query from the one or more related search queries, normalizing the weighted click data for the document and the particular related search query by dividing the weighted click data for the document and the particular related search query by a number of times the related search query was received over the period of time;

wherein combining the weighted click data comprises combining normalized weighted click data for the document and the one or more related search queries.

17. The system of claim 16, wherein combining the normalized weighted click data comprises summing the normalized weighted click data for the one or more related search queries.

18. The system of claim 15, wherein the operations further comprise combining additional click data with the combined weighted click data to generate the score, wherein the additional click data indicates how frequently the document was selected when the document was presented in search results for the search query.

19. The system of claim 18, wherein the combined weighted click data is adjusted by a blending factor before it is combined with the additional click data, wherein the blending factor indicates a degree to which the click data for the document and the one or more related search queries reliably indicates user behavior with regard to the search query.

20. The system of claim 19, wherein combining the additional click data with the combined weighted click data comprises adding i) the combined weighted click data multiplied by the blending factor and ii) the additional click data multiplied by one minus the blending factor.

21. The system of claim 15, wherein the click data corresponds to a number of times that the document was selected when it was presented in the search results for the one or more related search queries over the period of time.