Methods and systems for adjusting a scoring measure based on query breadth

• US 8,060,517 B2
• Filed: 04/08/2011
• Issued: 11/15/2011
• Est. Priority Date: 03/17/2004
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Pfleger, Karl Robert, Larson, Brian
• Primary Examiner(s): ABEL JALIL, NEVEEN
• Assistant Examiner(s): Pulliam; Christyann
• Application Number:
  US13/082,732
• Publication Number:
  US 20110184930A1
• Time in Patent Office: 221 Days
• Field of Search: None
• US Class Current: 707/751
• CPC Class Codes: G06F 16/334 Query execution G06F16/335 takes precedence

Abstract

Methods and systems for adjusting a scoring measure of a search result based at least in part on the breadth of a previously-executed search query associated with the search result are described. In one described system, a search engine determines a popularity measure for a search result, and then adjusts the popularity measure based at least in part on a query breadth measure of a previously-executed search query associated with the search result. The search engine may use a variety of query breadth measures. For example, the search engine may use the quantity of results returned by the search query, the length of the query, the IR score drop-off, or some other measure of breadth.

Claims

1. A method implemented by a system of one or more computer processors, the method comprising:

receiving, at the system, machine-readable information characterizing a first user selection of a first search result that was identified by a search engine as responsive to a first search query;

receiving, at the system, machine-readable information characterizing a second user selection of a second search result that was identified by a search engine as responsive to a second search query;

determining, by the system, that breadth of the first search query is larger than breadth of the second search query; and

recording, in response to the determination, a characterization of the second user selection of the second search result as more strongly indicative of popularity than the first user selection of the first search result, the characterization being recorded in a machine-readable popularity database that stores measures of popularity.

2. The method of claim 1, wherein recording the characterization of the second user selection as more strongly indicative of popularity than the first user selection comprises de-weighting a popularity measure corresponding to a total number of clicks on the first user selection after the first user selection is retrieved in response to the first search query.

3. The method of claim 1, wherein recording the characterization of the second user selection as more strongly indicative of popularity than the first user selection comprises:

adjusting a first popularity measure associated with the first search result for the first search query; and

adjusting a second popularity measure associated with the second search result for the second search query.

4. The method of claim 1, wherein recording the characterization of the second user selection as more strongly indicative of popularity than the first user selection comprises increasing a ranking of the second search result to embody the second user selection.

5. The method of claim 1, further comprising:

estimating, by the system, the breadth of the first search query; and

estimating, by the system, the breadth of the second search query.

6. The method of claim 5, wherein estimating the breadth of the first search query comprises estimating the breadth based on differences in relevance of documents in a result set that is responsive to the first search query.

7. The method of claim 5, wherein estimating the breadth of the first search query comprises estimating the breadth based on a number of terms within the first search query or a length of the first search query.

8. The method of claim 5, wherein estimating the breadth of the first search query comprises estimating a frequency at which the first search query has been issued by users.

9. The method of claim 1, wherein the machine-readable information characterizing the first and second user selections comprises a machine-readable log of search queries paired with results responsive to the search queries, the results in the record having been selected by human users who submitted the paired of the search queries.

10. A system comprising:

a popularity database that stores measures of popularity of electronic documents available on the Internet; and

one or more computer processors implementing a search engine programmed to

receive search queries,

identify electronic documents that are available on the Internet and that are responsive to the received search queries,

provide result sets characterizing the identified documents,

identify user selections of electronic documents from the result sets, and

adjust the measures of popularity of the selected electronic documents in the popularity database, the adjustments characterizing user selections of electronic documents from results sets provided in response to broader search queries as less strongly indicative of popularity than user selections of electronic documents from results sets provided in response to narrower search queries.

11. The system of claim 10, wherein the search engine is further programmed to use the adjusted measures of popularity to sort electronic documents that are available on the Internet and that are responsive to subsequently received search queries.

12. The system of claim 10, wherein the search engine comprises a query breadth analyzer programmed to determine breadths of search queries.

13. The system of claim 10, wherein:

the system further comprises a log file of records of search queries and identifiers of electronic documents that were selected; and

the query breadth analyzer is programmed to retrieve records from the log file.

14. The system of claim 10, wherein the system further comprises a client device in communication with the search engine.

15. The system of claim 10, wherein adjusting the measures of popularity comprises de-weighting selections corresponding to user selections of electronic documents from results sets provided in response to broader search queries.

16. The system of claim 10, wherein the measures of popularity each correspond to individual search queries.

17. A method implemented by a search engine system of one or more computer processors, the method comprising:

receiving, at the search engine system, search queries,

identifying, by the search engine system, electronic documents that are available on the Internet and that are responsive to the received search queries,

providing, by the search engine system, result sets characterizing the identified documents,

identifying, by the search engine system, user selections of electronic documents from the result sets, and

adjusting, by the search engine system, measures of popularity of the selected electronic documents in a popularity database, the adjustments characterizing user selections of electronic documents from results sets provided in response to broader search queries as less strongly indicative of popularity than user selections of electronic documents from results sets provided in response to narrower search queries.

18. The method of claim 17, wherein adjusting the measures of popularity comprises de-weighting selections corresponding to user selections of electronic documents from results sets provided in response to broader search queries.

19. The method of claim 17, wherein the measures of popularity each correspond to individual search queries.

20. The method of claim 17, wherein adjusting the measures of popularity comprises increasing a ranking of the second search result to embody the second user selection.

21. A system comprising:

a popularity database that stores measures of popularity of electronic documents available on the Internet; and

one or more computer processors implementing a query breadth analyzer programmed to

receive machine-readable information characterizing a first user selection of a first search result that was identified by a search engine as responsive to a first search query;

receive machine-readable information characterizing a second user selection of a second search result that was identified by a search engine as responsive to a second search query;

determine that breadth of the first search query is larger than breadth of the second search query; and

record a characterization of the second user selection of the second search result as more strongly indicative of popularity than the first user selection of the first search result, the characterization being recorded in a machine-readable popularity database that stores measures of popularity.

22. The system of claim 21, comprising one or more computer processors implementing a search engine programmed to use the characterization of the second user selection as more strongly indicative of popularity than the first user selection to sort electronic documents that are available on the Internet and that are responsive to subsequently received search queries.

23. The system of claim 21, further comprising a client device in communication with the query breadth analyzer.