Augmenting search query results with behaviorally related items

• US 7,412,442 B1
• Filed: 10/15/2004
• Issued: 08/12/2008
• Est. Priority Date: 10/15/2004
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Vadon, Eric R., Few, Jeffrey D.
• Primary Examiner(s): Wong; Leslie
• Application Number:
  US10/966,343
• Time in Patent Office: 1,397 Days
• Field of Search: 707/3, 707/5, 707/7, 707/101, 707/104.1
• US Class Current: 1/1
• CPC Class Codes: G06F 16/3322 using system suggestions G06F16/3325 takes precedence
• CPC Class Codes: G06F 16/3325 Reformulation based on results of preceding query
• CPC Class Codes: G06F 16/337 Profile generation, learning or modification
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access
• CPC Class Codes: Y10S 707/99937 Sorting

Abstract

The results of search queries are augmented with behaviorally related items. Items that are behaviorally related to a particular search query may be identified by analyzing the item selection actions of users who submit the particular search query. For example, if a relatively large number of users who submit a particular search query eventually select a particular item (e.g., during subsequent searching or browsing), that item may subsequently be included in the search results for the search query, even though the item may not match the search query.

Claims

1. A computer-implemented method of improving search results based on actions of users, the method comprising:

maintaining search event data and item selection event data in computer storage, the search event data descriptive of search queries submitted by users to a search engine to search for items, the item selection event data descriptive of item selections actions performed by users;

calculating, and storing in computer memory, a score that represents a degree to which a particular search query is related to a particular item, said item being an item that is not included in a query result of said search query, said score being dependent upon the search event data and the item selection event data, and being reflective of a frequency with which users who submit the search query subsequently select the item; and

based on said score, augmenting the query result of the search query with said item to cause the item to be presented to a user in search results for said search query.

2. The method of claim 1, wherein calculating the score comprises using a score generation method that produces a normalized score falling within a bounded range having an upper bound and a lower bound.

3. The method of claim 1, wherein calculating the score comprises augmenting the score in response to an event sequence in which a user:

(1) submits the search query and is presented with search results that do not include the item, (2) subsequently, submits an alternative search query and is presented with search results that include the item, and (3) selects the item from said search results of the alternative search query.

4. The method of claim 1, wherein calculating the score comprises taking into consideration (a) a number of users who both submitted the search query and selected the item, and (b) a total number of users who submitted the search query.

5. The method of claim 1, wherein calculating the score comprises giving different amounts of weight to different types of item selection actions.

6. The method of claim 1, wherein the item is a product represented in an electronic catalog.

7. A system that provides improved search results to users, the system comprising:

a computer data repository that stores search event data and item selection event data, the search event data descriptive of search queries submitted by users to a search engine to search for items, the item selection event data descriptive of item selection actions performed by users;

a computer system programmed to use the search event data and the item selection event data to calculate a score that represents a degree to which a particular search query is related to a particular item, said item being an item that is not included in a query result of said search query, said score reflecting a tendency of users who submit the search query to subsequently select the item, said computer system comprising a memory; and

a search engine system programmed to use the score to determine whether to augment the query result of the search query with said item.

8. The system of claim 7, wherein the computer system is programmed to calculate the score using a score generation method that produces a normalized score falling within a bounded range having an upper bound and a lower bound.

9. The system of claim 7, wherein the computer system is programmed to augment the score in response to an event sequence in which a user:

(1) submits the search query and is presented with search results that do not include the item, (2) subsequently, submits an alternative search query and is presented with search results that include the item, and (3) selects the item from said search results of the alternative search query.

10. The system of claim 7, wherein the computer system is programmed to give different amounts of weight to different types of item selection actions.

11. The system of claim 7, wherein the computer system is programmed to generate the score using a method that takes into consideration (a) a number of users who both submitted the search query and selected the item, and (b) a total number of users who submitted the search query.

12. The system of claim 7, wherein the item is a catalog item represented on an item detail page of an electronic catalog, and the computer system is programmed to calculate the score using a method that gives greater weight to a purchase of the item than to a viewing event in which a user views said item detail page.