Method for personalized search

• US 20050102282A1
• Filed: 10/12/2004
• Published: 05/12/2005
• Est. Priority Date: 11/07/2003
• Status: Abandoned Application

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Greg Linden
• Inventors: Linden, Greg
• Application Number:
  US10/961,974
• Publication Number:
  US 20050102282A1
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/24578 using ranking
• CPC Class Codes: G06F 16/284 Relational databases
• CPC Class Codes: G06F 16/90324 using system suggestions

Abstract

A search tool provides a means of finding a set of items in a large collection of items using a search query. Personalized search generates different search results to different users of the search engine based on their interests and past behavior. The invention describes a method of providing personalized search using previous search queries of the user, pages viewed from previous search results, and the pages viewed by other users with similar searches.

Claims

1. In a multi-user computer system that provides user access to a database of items, a method of providing personalized search results from the database, the method comprising the computer-implemented steps of:

(a) generating a data structure which maps individual search queries in a database to corresponding sets of similar queries where similarity is based at least in part upon correlations between queries made by users of the search engine;

(b) generating a data structure which maps individual search result items in a database to corresponding sets of similar items in which similarities between items are based at least in part upon correlations between items viewed by users of the search engine;

(c) for a search query, accessing the data structure in step (a) to identify a corresponding set of similar queries;

(d) for search result items, accessing the data structure in step (b) to identify a corresponding set of similar search result items; and

(e) modifying search results for a given search query based at least in part on similar queries and similar search result items;

wherein step (a)-(b) is performed in an off-line mode, and steps (c)-(e) are performed substantially in real time in response to an online action by the user.

2. The method of claim 1, wherein step (e) comprises of emphasizing search results items frequently viewed by other users on similar search queries.

3. The method of claim 1, wherein step (e) comprises of deemphasizing search result items previously shown to the user for similar search queries.

4. The method of claim 1, wherein step (e) comprises of emphasizing search result items that are similar to search result items viewed by the user on previous search queries that are similar to the current search query.

5. A method of modifying results from a database of items comprised the computer-implemented steps of:

(a) accessing the database using a search query;

(b) accessing a database containing a history of queries and search results viewed by the user;

(c) accessing a database containing similar search queries for any given search query;

(d) accessing a database containing the most popular search result items for any given search query;

(e) accessing a database containing similar search result items for any given search result item;

(f) modifying the search results produced in step (a) using the set from step (b);

(g) modifying the search results produced in step (a) using the set from step (c);

(h) modifying the search results produced in step (a) using the set from step (d);

(i) modifying the search results produced in step (a) using the set from step (e);

(j) combining the modified search results from steps (f)-(i).

6. The method of claim 5, wherein the database in step (a) is a web-based search engine.

7. The method of claim 5, wherein step (b) is an in-memory database containing a finite history of the queries and search results for the queries.

8. The method of claim 5, wherein the database in step (c) is built from the history of user'"'"'s searches on the database.

9. The method of claim 5, wherein the database in step (c) is built at least in part by analyzing correlations between search queries made by users of the search engine.

10. The method of claim 5, wherein the database in step (e) is built at least in part by analyzing correlations between search result items viewed by users of the search engine.

11. The method of claim 5, wherein steps (f) and (g) reduce the rank of search result items previously seen by the user for the same or similar search queries.

12. The method of claim 5, wherein step (h) increases the rank of search result items popular with other users making similar search queries.

13. The method of claim 5, wherein step (i) increases the rank of search result items that are similar to search result items previously viewed by the user for the same or similar search queries.

14. A method of searching a database of items where the search results are modified based on previous similar search queries, the method comprising of:

(a) finding similar search queries at least in part by analyzing correlations between the searches of users of the search engine;

(b) increasing the rank of search result items for the current search query that were frequently viewed by other users of the search engine when they executed a search query similar to the current user'"'"'s search query.

15. A method of searching a database of items where the search results are modified based on previous similar search queries, the method comprising of:

(a) finding similar search queries at least in part by analyzing correlations between the searches of users of the search engine;

(b) decreasing the rank of search result items for the current search query that were previously seen by the user on similar search queries.

16. A method of searching a database of items where the search results are modified based on similarities between search result items, the method comprising of:

(a) finding similar search result items at least in part by analyzing correlations between the search result items viewed by users of the search engine;

(b) finding similar search queries at least in part by analyzing correlations between the searches of users of the search engine;

(c) increasing the rank of a search result items for the current search query that are similar to a search result item previously viewed by the user on the same or a similar search query.