Refinement surfacing for search hierarchies

• US 8,694,529 B1
• Filed: 01/28/2010
• Issued: 04/08/2014
• Est. Priority Date: 01/28/2010
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Hotchkies, Blair L., Parshall, Daniel R., Patrikios, Jason P., Mills, Brent Robert, Teodorescu, Dan Catalin, Pounds, Jennica Jane, Jaye, Paul Daniel, Lichterman, David Aaron
• Primary Examiner(s): Wu, Yicun
• Application Number:
  US12/695,990
• Time in Patent Office: 1,531 Days
• Field of Search: 707/766, 707/767
• US Class Current: 707/766
• CPC Class Codes: G06F 16/3349 Reuse of stored results of previous queries
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 16/9538 Presentation of query results

Abstract

A collection of items may be categorized according to a category hierarchy. A search of the items may be explicitly associated with a particular category in the hierarchy or else unassociated. A set of refinements may be available, for example, to help narrow the search. Refinements may be applicable to items in linked categories. Refinements may be surfaced based at least in part on linked categories associated with the search. Refinements may be surfaced when the search is unassociated with a particular category. For example, a possibly relevant set of categories may be determined. Refinements may be surfaced according to a variety of ordering criteria.

Claims

1. A computerized system for facilitating a search of a plurality of items, comprising:

a data store storing, at least;

a hierarchy of categories that categorize at least some of the plurality of items;

a set of search refinements, at least some of the hierarchy of categories being linked to at least some of the set of search refinements; and

a refinement surfacing map depicting a plurality of associations between search terms and at least some of the hierarchy of categories;

a search module configured at least to

generate a search result set of items from the plurality of items that match one or more search terms; and

maintain a log of a number of times a refinement of the set of search refinements is selected during searches employing the one or more search terms in one or more time periods;

a refinement surfacing module configured to, at least;

determine a possibly relevant set of categories of the hierarchy of categories below a distinguished category based at least in part on the log;

increase an associative confidence between at least one of the one or more search terms and a subset of the set of search refinements linked to the possibly relevant set of categories based at least in part on the log; and

determine the subset of the set of search refinements linked to the possibly relevant set of categories when the associative confidence at least meets a first sufficiency threshold;

a user interface module configured at least to, as least

provide the search result set of items and said subset of the set of search refinements for presentation;

constrain subsequent search result sets to a particular distinguished category, based at least on the one or more search terms, when the associative confidence meets at least a second sufficiency threshold; and

one or more processors collectively facilitating at least the data store, the search module, the refinement surfacing module and the user interface module.

2. A computerized system according to claim 1, wherein the distinguished category is a root category of the hierarchy of categories.

3. A computerized system according to claim 1, wherein generating the search result set of items from the plurality of items that match one or more search terms comprises:

for each of the plurality of items in the search result set of items, determining a relevance score for one of the plurality of items with respect to the one or more search terms; and

determining whether the relevance score for the one of the plurality of items is above a threshold.

4. A computer-implemented method for facilitating a search of a plurality of items categorized according to a category hierarchy, comprising:

under control of one or more computer systems configured with executable instructions,

generating a search result set of items from the plurality of items that match one or more search terms;

identifying at least one category in the category hierarchy below a distinguished category based at least in part on the one or more search terms and a map depicting a plurality of associations between search terms and at least some categories of the category hierarchy;

receiving a selection of at least one refinement from a set of refinements linked to said at least one category;

maintaining a log of a number of times the at least one refinement from the set of refinements is selected during searches employing the one or more search terms in one or more time periods;

determining a possibly relevant set of categories of the category of hierarchy below a distinguished category based at least in part on the log;

increasing an associative confidence between at least one of the one or more search terms and a subset of the set of refinements linked to the possibly relevant set of categories based at least in part on the log;

determining the subset of the set of refinements linked to the possibly relevant set of categories when the associative confidence at least meets a first sufficiency threshold;

providing the search result set of items and said subset of the set of refinements for presentation; and

constraining subsequent search result sets to a particular distinguished category, based at least on the one or more search terms, when the associative confidence meets at least a second sufficiency threshold.

5. A computer-implemented method according to claim 4, further comprising ordering the set of refinements in accordance with at least one refinement surfacing policy.

6. A computer-implemented method according to claim 4, wherein selecting said at least one refinement from the set of refinements comprises selecting said at least one refinement from the set of refinements in an order specified by the ordering of the set of refinements.

7. A computer-implemented method for facilitating a search, comprising:

under control of one or more computer systems configured with executable instructions,

associating the search with a category in a category hierarchy;

searching the category to find items matching one or more search terms;

identifying at least one further category at a level below the category in the category hierarchy based at least in part on the one or more search terms and a map for surfacing refinements, the map depicting a plurality of associations between search terms and the at least one further category;

receiving a selection of at least one refinement from a set of refinements linked to said at least one further category;

maintaining a log of a number of times a refinement of the set of refinements is selected during searches employing the one or more search terms in one or more time periods;

determining a possibly relevant set of categories of the category hierarchy below a distinguished category based at least in part on the log;

increasing an associative confidence between at least one of the one or more search terms and a subset of the set of refinements linked to the possibly relevant set of categories based at least in part on the log;

determining the subset of the set of refinements linked to the possibly relevant set of categories when the associative confidence at least meets a first sufficiency threshold;

providing the items and said subset of the set of refinements for presentation; and

constraining subsequent search results to a particular distinguished category, based at least on the one or more search terms, when the associative confidence meets at least a second sufficiency threshold.

8. A computer-implemented method according to claim 7, wherein:

the category hierarchy has a root category; and

the category associated with the search is the root category of the category hierarchy.

9. A computer-implemented method according to claim 7, wherein:

the map for surfacing refinements comprises a map from search terms to at least some categories in the category hierarchy; and

identifying said at least one further category comprises applying the map to the one or more search terms.

10. A computer-implemented method according to claim 7, wherein a refinement is linked to a category when the refinement is applicable to items in the category.

11. A computer-implemented method according to claim 7, further comprising:

determining a containing sub-hierarchy that includes each category linked to a distinguished refinement of said at least one refinement; and

associating the search with a root category of the containing sub-hierarchy.

12. A computer-implemented method for facilitating a search, comprising:

under control of one or more computer systems configured with executable instructions,

categorizing at least some of a first plurality of items included in a search result set of items into at least one of a plurality of categories arranged in a category hierarchy having a root category;

selecting a second plurality of items from the first plurality of items based at least in part on one or more search terms and independent of categories in the category hierarchy other than the root category;

selecting a plurality of refinements from a set of refinements based at least in part on the one or more search terms and a map depicting a plurality of associations between search terms and categories of the category hierarchy, and links between the categories and the set of refinements;

receiving a selection of at least one of the plurality of refinements linked to the at least one of a plurality of categories;

maintaining a log of a number of times the at least one refinement of the plurality of refinements is selected during searches employing the one or more search terms in one or more time periods;

determining a possibly relevant set of categories of the category hierarchy below a distinguished category based at least in part on the log;

increasing an associative confidence between at least one of the one or more search terms and a subset of the set of refinements linked to the possibly relevant set of categories based at least in part on the log;

determining the subset of the set of refinements linked to the possibly relevant set of categories when the associative confidence at least meets a first sufficiency threshold;

providing said subset of the set of refinements for presentation; and

constraining subsequent search result sets to a particular distinguished category, based at least on the one or more search terms, when the associative confidence meets at least a sufficiency second threshold.

13. A computer-implemented method according to claim 12, wherein selecting the plurality of refinements comprises:

identifying a possibly relevant set of categories by applying the map to the one or more search terms; and

selecting refinements in the set of refinements linked to categories in the possibly relevant set of categories.

14. A computer-implemented method according to claim 12, wherein ordering the plurality of refinements comprises determining a relevance score for at least some of the plurality of refinements.

15. A computer-implemented method according to claim 14, wherein determining the relevance score for each of the plurality of refinements comprises determining an efficiency of each of the plurality of refinements.

16. A computer-implemented method according to claim 14, wherein determining the relevance score for each of the plurality of refinements comprises determining a commercial efficiency of each of the plurality of refinements.

17. A computer-implemented method according to claim 12, wherein ordering the plurality of refinements comprises assigning a weight to at least one of the plurality of refinements of the plurality of refinements based at least in part on a distance in the category hierarchy from a root of at least one linked category.

18. A computer-implemented method according to claim 12, wherein ordering the plurality of refinements comprises determining a popularity score for at least some of the plurality of refinements.

19. A computer-implemented method according to claim 18, wherein the popularity score for each of the plurality of refinements is based at least in part on a number of times a refinement was historically selected during searches employing the one or more search terms in one or more time periods.

20. A computer-implemented method according to claim 12, wherein ordering the plurality of refinements comprises ordering at least some of the plurality of refinements pseudorandomly.

21. A computer-implemented method according to claim 12, wherein ordering the plurality of refinements comprises ordering at least some of the plurality of refinements according to a specified ordering.

22. One or more non-transitory computer-readable media having collectively thereon computer-executable instructions that configure one or more computers to collectively, at least:

associate a search with a category in a category hierarchy;

search the category to find items matching one or more search terms;

identify at least one further category below the category in the category hierarchy based at least in part on the one or more search terms and a map for surfacing refinements, the map depicting a plurality of associations between search terms and the at least one further category;

receive a selection of at least one refinement from a set of refinements linked to said at least one further category;

maintain a log of a number of times a refinement of the set of refinements is selected during searches employing the one or more search terms in one or more time periods;

determine a possibly relevant set of categories of the category hierarchy below a distinguished category based at least in part on the log;

increase an associative confidence between at least one of the one or more search terms and a subset of the set of refinements linked to the possibly relevant set of categories based at least in part on the log;

determine the subset of the set of refinements linked to the possibly relevant set of categories when the associative confidence at least meets a first sufficiency threshold;

provide the items and said subset of the set of refinements for presentation; and

,

constrain subsequent search results to a particular distinguished category, based at least on the one or more search terms, when the associative confidence meets at least a second sufficiency threshold.

23. One or more non-transitory computer-readable media according to claim 22, wherein the set of refinements comprises a hierarchy of refinements.

24. One or more non-transitory computer-readable media according to claim 23, wherein selecting said at least one refinement is further based at least in part on the hierarchy.

25. One or more non-transitory computer-readable media according to claim 22, wherein said at least one refinement comprises refinements below a distinguished refinement in a hierarchy of refinements.