Identifying navigational resources for informational queries

• US 9,244,972 B1
• Filed: 12/31/2012
• Issued: 01/26/2016
• Est. Priority Date: 04/20/2012
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Lehman, April R., Fernandes, Neil C., Chakrapani, Lakshmi N.
• Primary Examiner(s): Liao, Jason
• Application Number:
  US13/731,857
• Time in Patent Office: 1,121 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06F 16/245 Query processing
• CPC Class Codes: G06F 16/3326 using relevance feedback from the user, e.g. relevance feedback on documents, documents sets, document terms or passages
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation

Abstract

Methods, systems, and apparatus, including computer program products, for creating a mapping between topics and navigational resources are disclosed. The mapping is created based on hybrid queries having both characteristics of informational queries and navigational queries. The mapping between topics and resources can be used in scoring search results for ranking, for example. Other mappings created based on other types of hybrid queries and their applications are also disclosed.

Claims

1. A method performed by one or more computers, the method comprising:

processing a query log to identify a plurality of hybrid queries, each hybrid query including a respective topic keyword that identifies a respective topic and a respective navigation keyword that identifies a respective navigational resource;

creating a first mapping, the first mapping including a plurality of first associations each associating one of a plurality of unique navigational resources among the respective navigational resources identified in the plurality of hybrid queries with a respective group of one or more topics that are each identified in at least one of the hybrid queries that also identify the unique navigational resource;

for each of one or more of the first associations in the first mapping;

generalizing the respective group of one or more topics in the first association to identify one or more additional topics not present in the respective group; and

augmenting the first association to associate the one or more additional topics with the unique navigational resource in the first association;

inverting the first mapping to obtain a second mapping, the second mapping including a plurality of second associations each associating one of a plurality of unique topics among the topics present in the first mapping with a respective group of one or more navigational resources, where each of the one or more navigational resources of the respective group is associated with the unique topic in the first mapping;

detecting that terms of a partially-completed search query received from a search input interface match a first unique topic present in the second mapping;

identifying one or more navigational resources associated with the first unique topic; and

scoring candidate auto-completions for the partially-completed search query based at least on whether the candidate auto-completions have search results leading to the one or more navigational resources that have been identified.

2. The method of claim 1, wherein creating the first mapping comprises:

filtering a plurality of initial associations based on one or more association precision criteria, each initial association associating a unique navigational resource among the respective navigational resources identified in the plurality of hybrid queries and a unique topic among the respective topics identified in one or more of the plurality of hybrid queries that identify the unique navigational resource;

based on the filtered initial associations, generating an initial topic-to-resource mapping, the initial topic-to-resource mapping including a plurality of third associations each associating one of a plurality of unique topics among the topics present in the filtered initial associations with a respective group of one or more navigational resources, where each navigational resource of the respective group is associated with the unique topic in a respective one of the filtered initial associations; and

inverting the initial topic-to-resource mapping to generate the first mapping.

3. The method of claim 2, wherein:

the one or more association precision criteria include a minimum fraction that a click count for the unique navigational resource identified in each filtered initial association accounts for in a total click count for search results of the hybrid queries that identify both a unique navigational resource in the plurality of unique navigational resources and the unique topic of the filtered initial association.

4. The method of claim 2, wherein:

the one or more association precision criteria include a minimum number of recorded search sessions in which query revisions from an informational query for the unique topic of each filtered initial association to a navigational query for the unique navigational resource of the filtered initial association have occurred.

5. The method of claim 2, wherein:

the one or more association precision criteria include a requirement that the unique navigational resource of each filtered initial association receives at least a threshold number of recorded clicks for informational queries containing the respective topic keyword of the unique topic identified in the filtered initial association.

6. The method of claim 1, further comprising:

for each of one or more of the second associations in the second mapping, calculating a respective association score for at least one navigational resource identified in the second association,

where using the second mapping in scoring the candidate search results for the search queries comprises using the respective association scores of the navigational resources in the one or more of the second associations in scoring the candidate search results for the search queries.

7. The method of claim 6, wherein calculating the respective association score for the at least one navigational resource identified in the second association further comprises:

for a first navigational resource identified in the second association;

identifying, from the query log, a plurality of informational queries consisting of the topic keyword of the unique topic in the second association;

identifying one or more of the informational queries for which the first navigational resource identified in the second association has received one or more clicks as a search result; and

calculating the respective association score for the first navigational resource based at least on a total number of clicks that the first navigational resource identified in the second association has received as the search result for the one or more of the informational queries.

8. The method of claim 6, wherein calculating the association score for the at least one navigational resource identified in the second association further comprises:

for a first navigational resource identified in the second association;

determining a count of recorded search sessions in which query revisions from an informational query for the unique topic identified in the second association to a navigational query for the first navigational resource identified in the second association have occurred; and

calculating the association score for the first navigational resource based at least on the count of the recorded search sessions that has been determined.

9. The method of claim 1, wherein calculating the association score for the at least one navigational resource identified in the second association further comprises:

for a first navigational resource identified in the second association;

determining, among the second associations in the second mapping, a total count of the unique topics associated with the first navigational resource; and

calculating the association score for the first navigational resource based at least on the total count of the unique topics.

10. The method of claim 1, wherein calculating the association score for the at least one navigational resource identified in the second association further comprises:

for a first navigational resource identified in the second association;

identifying, from a document corpus, respective anchor text of a plurality of hyperlinks pointing to the first navigational resource identified in the second association;

determining a count of appearances that the topic keyword of the unique topic in the second association makes in the respective anchor text of the plurality of hyperlinks; and

calculating the association score for the first navigational resource based at least on the count of appearances.

11. The method of claim 1, further comprising:

categorizing each unique topic in the second mapping into respective one or more of a plurality of unique verticals;

creating a vertical-to-resource mapping comprising a plurality of vertical-to-resource associations, each vertical-to-resource association associating a respective one of the unique verticals with a respective group of navigational resources, where each of the navigational resources of the respective group is associated with at least one of the unique topics in the second mapping that has been categorized into the respective unique vertical; and

using the vertical-to-resource mapping in scoring candidate search results for search queries that are categorized into one or more of the plurality of unique verticals.

12. The method of claim 1, wherein the navigational resources in the second mapping are websites or webpages having respective associated web addresses.

13. The method of claim 1, further comprising:

determining a first topic related to a user-submitted search query;

identifying one or more navigational resources associated with the first topic according to the second mapping; and

providing a user interface element on a search result webpage for the user-submitted search query, the user interface element configured to identify the one or more navigational resources associated with the first topic.

14. A non-transitory medium storing instructions that, when executed by one or more computers, cause the one or more computers to perform operations comprising:

processing a query log to identify a plurality of hybrid queries, each hybrid query including a respective topic keyword that identifies a respective topic and a respective navigation keyword that identifies a respective navigational resource;

creating a first mapping, the first mapping including a plurality of first associations each associating one of a plurality of unique navigational resources among the respective navigational resources identified in the plurality of hybrid queries with a respective group of one or more topics that are each identified in at least one of the hybrid queries that also identify the unique navigational resource;

for each of one or more of the first associations in the first mapping;

generalizing the respective group of one or more topics in the first association to identify one or more additional topics not present in the respective group; and

augmenting the first association to associate the one or more additional topics with the unique navigational resource in the first association;

inverting the first mapping to obtain a second mapping, the second mapping including a plurality of second associations each associating one of a plurality of unique topics among the topics present in the first mapping with a respective group of one or more navigational resources, where each of the one or more navigational resources of the respective group is associated with the unique topic in the first mapping;

detecting that terms of a partially-completed search query received from a search input interface match a first unique topic present in the second mapping;

identifying one or more navigational resources associated with the first unique topic; and

scoring candidate auto-completions for the partially-completed search query based at least on whether the candidate auto-completions have search results leading to the one or more navigational resources that have been identified.

15. The medium of claim 14, wherein creating the first mapping comprises:

filtering a plurality of initial associations based on one or more association precision criteria, each initial association associating a unique navigational resource among the respective navigational resources identified in the plurality of hybrid queries and a unique topic among the respective topics identified in one or more of the plurality of hybrid queries that identify the unique navigational resource;

based on the filtered initial associations, generating an initial topic-to-resource mapping, the initial topic-to-resource mapping including a plurality of third associations each associating one of a plurality of unique topics among the topics present in the filtered initial associations with a respective group of one or more navigational resources, where each navigational resource of the respective group is associated with the unique topic in a respective one of the filtered initial associations; and

inverting the initial topic-to-resource mapping to generate the first mapping.

16. The medium of claim 15, wherein:

the one or more association precision criteria include a minimum fraction that a click count for the unique navigational resource identified in each filtered initial association accounts for in a total click count search results of the hybrid queries that identify both a unique navigational resource in the plurality of unique navigational resources and the unique topic of the filtered initial association.

17. The medium of claim 15, wherein:

the one or more association precision criteria include a minimum number of recorded search sessions in which query revisions from an informational query for the unique topic of each filtered initial association to a navigational query for the unique navigational resource of the filtered initial association have occurred.

18. The medium of claim 15, wherein:

the one or more association precision criteria include a requirement that the unique navigational resource of each filtered initial association receives at least a threshold number of recorded clicks for informational queries containing the respective topic keyword of the unique topic identified in the filtered initial association.

19. The medium of claim 14, wherein the operations further comprise:

for each of one or more of the second associations in the second mapping, calculating a respective association score for at least one navigational resource identified in the second association,

where using the second mapping in scoring the candidate search results for the search queries comprises using the respective association scores of the navigational resources in the one or more of the second associations in scoring the candidate search results for the search queries.

20. The medium of claim 19, wherein calculating the respective association score for the at least one navigational resource identified in the second association further comprises:

for a first navigational resource identified in the second association;

identifying, from the query log, a plurality of informational queries consisting of the topic keyword of the unique topic in the second association;

identifying one or more of the informational queries for which the first navigational resource identified in the second association has received one or more clicks as a search result; and

calculating the respective association score for the first navigational resource based at least on a total number of clicks that the first navigational resource identified in the second association has received as the search result for the one or more of the informational queries.

21. The medium of claim 19, wherein calculating the association score for the at least one navigational resource identified in the second association further comprises:

for a first navigational resource identified in the second association;

determining a count of recorded search sessions in which query revisions from an informational query for the unique topic identified in the second association to a navigational query for the first navigational resource identified in the second association have occurred; and

calculating the association score for the first navigational resource based at least on the count of the recorded search sessions that has been determined.

22. The medium of claim 14, wherein calculating the association score for the at least one navigational resource identified in the second association further comprises:

for a first navigational resource identified in the second association;

determining, among the second associations in the second mapping, a total count of the unique topics associated with the first navigational resource; and

calculating the association score for the first navigational resource based at least on the total count of the unique topics.

23. The medium of claim 14, wherein calculating the association score for the at least one navigational resource identified in the second association further comprises:

for a first navigational resource identified in the second association;

identifying, from a document corpus, respective anchor text of a plurality of hyperlinks pointing to the first navigational resource identified in the second association;

determining a count of appearances that the topic keyword of the unique topic in the second association makes in the respective anchor text of the plurality of hyperlinks; and

calculating the association score for the first navigational resource based at least on the count of appearances.

24. The medium of claim 14, wherein the operations further comprise:

categorizing each unique topic in the second mapping into respective one or more of a plurality of unique verticals;

creating a vertical-to-resource mapping comprising a plurality of vertical-to-resource associations, each vertical-to-resource association associating a respective one of the unique verticals with a respective group of navigational resources, where each of the navigational resources of the respective group is associated with at least one of the unique topics in the second mapping that has been categorized into the respective unique vertical; and

using the vertical-to-resource mapping in scoring candidate search results for search queries that are categorized into one or more of the plurality of unique verticals.

25. The medium of claim 14, wherein the navigational resources in the second mapping are websites or webpages having respective associated web addresses.

26. The medium of claim 14, wherein the operations further comprise:

determining a first topic related to a user-submitted search query;

identifying one or more navigational resources associated with the first topic according to the second mapping; and

providing a user interface element on a search result webpage for the user-submitted search query, the user interface element configured to identify the one or more navigational resources associated with the first topic.

27. A system, comprising:

one or more computers; and

memory having instructions stored thereon, the instructions, when executed by the one or more computers, cause the one or more computers to perform operations comprising;

processing a query log to identify a plurality of hybrid queries, each hybrid query including a respective topic keyword that identifies a respective topic and a respective navigation keyword that identifies a respective navigational resource;

creating a first mapping, the first mapping including a plurality of first associations each associating one of a plurality of unique navigational resources among the respective navigational resources identified in the plurality of hybrid queries with a respective group of one or more topics that are each identified in at least one of the hybrid queries that also identify the unique navigational resource;

for each of one or more of the first associations in the first mapping;

generalizing the respective group of one or more topics in the first association to identify one or more additional topics not present in the respective group; and

augmenting the first association to associate the one or more additional topics with the unique navigational resource in the first association;

inverting the first mapping to obtain a second mapping, the second mapping including a plurality of second associations each associating one of a plurality of unique topics among the topics present in the first mapping with a respective group of one or more navigational resources, where each of the one or more navigational resources of the respective group is associated with the unique topic in the first mapping;

detecting that terms of a partially-completed search query received from a search input interface match a first unique topic present in the second mapping;

identifying one or more navigational resources associated with the first unique topic; and

scoring candidate auto-completions for the partially-completed search query based at least on whether the candidate auto-completions have search results leading to the one or more navigational resources that have been identified.

28. The system of claim 27, wherein creating the first mapping comprises:

filtering a plurality of initial associations based on one or more association precision criteria, each initial association associating a unique navigational resource among the respective navigational resources identified in the plurality of hybrid queries and a unique topic among the respective topics identified in one or more of the plurality of hybrid queries that identify the unique navigational resource;

based on the filtered initial associations, generating an initial topic-to-resource mapping, the initial topic-to-resource mapping including a plurality of third associations each associating one of a plurality of unique topics among the topics present in the filtered initial associations with a respective group of one or more navigational resources, where each navigational resource of the respective group is associated with the unique topic in a respective one of the filtered initial associations; and

inverting the initial topic-to-resource mapping to generate the first mapping.

29. The system of claim 28, wherein:

the one or more association precision criteria include a minimum fraction that a click count for the unique navigational resource identified in each filtered initial association accounts for in a total click count for search results of the hybrid queries that identify both a unique navigational resource in the plurality of unique navigational resources and the unique topic of the filtered initial association.

30. The system of claim 28, wherein:

the one or more association precision criteria include a minimum number of recorded search sessions in which query revisions from an informational query for the unique topic of each filtered initial association to a navigational query for the unique navigational resource of the filtered initial association have occurred.

31. The system of claim 30, wherein:

the one or more association precision criteria include a requirement that the unique navigational resource of each filtered initial association receives at least a threshold number of recorded clicks for informational queries containing the respective topic keyword of the unique topic identified in the filtered initial association.

32. The system of claim 27, wherein the operations further comprise:

for each of one or more of the second associations in the second mapping, calculating a respective association score for at least one navigational resource identified in the second association,

where using the second mapping in scoring the candidate search results for the search queries comprises using the respective association scores of the navigational resources in the one or more of the second associations in scoring the candidate search results for the search queries.

33. The system of claim 32, wherein calculating the respective association score for the at least one navigational resource identified in the second association further comprises:

for a first navigational resource identified in the second association;

identifying, from the query log, a plurality of informational queries consisting of the topic keyword of the unique topic in the second association;

identifying one or more of the informational queries for which the first navigational resource identified in the second association has received one or more clicks as a search result; and

calculating the respective association score for the first navigational resource based at least on a total number of clicks that the first navigational resource identified in the second association has received as the search result for the one or more of the informational queries.

34. The system of claim 32, wherein calculating the association score for the at least one navigational resource identified in the second association further comprises:

for a first navigational resource identified in the second association;

determining a count of recorded search sessions in which query revisions from an informational query for the unique topic identified in the second association to a navigational query for the first navigational resource identified in the second association have occurred; and

calculating the association score for the first navigational resource based at least on the count of the recorded search sessions that has been determined.

35. The system of claim 27, wherein calculating the association score for the at least one navigational resource identified in the second association further comprises:

for a first navigational resource identified in the second association;

determining, among the second associations in the second mapping, a total count of the unique topics associated with the first navigational resource; and

calculating the association score for the first navigational resource based at least on the total count of the unique topics.

36. The system of claim 27, wherein calculating the association score for the at least one navigational resource identified in the second association further comprises:

for a first navigational resource identified in the second association;

identifying, from a document corpus, respective anchor text of a plurality of hyperlinks pointing to the first navigational resource identified in the second association;

determining a count of appearances that the topic keyword of the unique topic in the second association makes in the respective anchor text of the plurality of hyperlinks; and

calculating the association score for the first navigational resource based at least on the count of appearances.

37. The system of claim 27, wherein the operations further comprise:

categorizing each unique topic in the second mapping into respective one or more of a plurality of unique verticals;

creating a vertical-to-resource mapping comprising a plurality of vertical-to-resource associations, each vertical-to-resource association associating a respective one of the unique verticals with a respective group of navigational resources, where each of the navigational resources of the respective group is associated with at least one of the unique topics in the second mapping that has been categorized into the respective unique vertical; and

using the vertical-to-resource mapping in scoring candidate search results for search queries that are categorized into one or more of the plurality of unique verticals.

38. The system of claim 27, wherein the navigational resources in the second mapping are websites or webpages having respective associated web addresses.

39. The system of claim 27, wherein the operations further comprise:

determining a first topic related to a user-submitted search query;

identifying one or more navigational resources associated with the first topic according to the second mapping; and

providing a user interface element on a search result webpage for the user-submitted search query, the user interface element configured to identify the one or more navigational resources associated with the first topic.