IDENTIFYING RELATED SEARCH TERMS BASED ON SEARCH BEHAVIORS OF USERS

• US 20110035370A1
• Filed: 10/19/2010
• Published: 02/10/2011
• Est. Priority Date: 07/15/1998
• Status: Active Grant

• Current Assignee: Dwayne E. Bowman, Joel R. Spiegel, Michael L. Hamrick, Ruben E. Ortega, Timothy R. Kohn
• Original Assignee: Dwayne E. Bowman, Joel R. Spiegel, Michael L. Hamrick, Ruben E. Ortega, Timothy R. Kohn
• Inventors: Hamrick, Michael L., Kohn, Timothy R., Bowman, Dwayne E., Ortega, Ruben E., Spiegel, Joel R.
• Granted Patent: US 7,996,398 B2
• Time in Patent Office: Days
• US Class Current: 707/721
• CPC Class Codes: G06F 16/3322 using system suggestions G06F16/3325 takes precedence
• CPC Class Codes: G06F 16/3338 Query expansion
• CPC Class Codes: G06F 16/90 Details of database functions independent of the retrieved data types
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 16/9532 Query formulation
• CPC Class Codes: G06F 40/226 Validation
• CPC Class Codes: G06F 40/232 Orthographic correction, e.g. spell checking or vowelisation
• CPC Class Codes: G06F 40/274 Converting codes to words; Guess-ahead of partial word inputs
• CPC Class Codes: Y10S 707/99932 Access augmentation or optimizing
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching
• CPC Class Codes: Y10S 707/99934 Query formulation, input preparation, or translation
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access
• CPC Class Codes: Y10S 707/99936 Pattern matching access
• CPC Class Codes: Y10S 707/99943 Generating database or data structure, e.g. via user interface

Abstract

A system is disclosed that identifies related search terms by performing a term co-occurrence analysis of search query submissions of users. The system may give different amounts of weight to different search query submissions based, for example, on whether subsequent user actions reveal that particular searches were successful. The system may use the identified related search terms to correct spelling errors in search queries received from users.

Claims

1. A computer-implemented method of identifying related terms, the method comprising:

analyzing search query submissions of a plurality of search engine users to measure degrees to which particular search terms are related to each other, said degrees of relationship measured based at least partly on how frequently particular search terms co-occur in multiple-term search queries of users, wherein analyzing the search query submissions comprises weighting some search query submissions differently than other search query submissions such that some search query submissions are given different non-zero amounts of weight than other search query submissions in measuring said degrees to which particular search terms are related; and

generating, in physical computer storage, a table reflective of a result of the analysis of search query submissions of the plurality of users, said table identifying specific search terms that are related to each other;

said method performed in its entirety by a computer system that comprises one or more computers.

2. The method of claim 1, wherein weighting some search query submissions differently than others comprises according, to a search query submission of a user, a weight that depends on whether the user viewed an associated query result item.

3. The method of claim 1, wherein weighting some search query submissions differently than others comprises giving greater weight to search query submissions that resulted in a user transaction than to search query submissions that did not result in a user transaction.

4. The method as in claim 1, further comprising, by a computer system, automatically using the table to predict correct spellings of selected search terms in multiple-term search queries submitted by users.

5. The method as in claim 1, further comprising using the table to programmatically analyze spellings of selected search terms in multiple-term search queries submitted by users.

6. The method as in claim 1, further comprising, via a computer system, automatically using the table to select a term to use in place of a non-matching search term in a multi-term search query submitted to a search engine by a current user.

7. The method as in claim 1, wherein analyzing the search query submissions comprises keeping track of specific search fields in which search terms were entered by users.

8. The method as in claim 1, wherein generating the table comprises incorporating, into the table, term correlation data derived from a term co-occurrence analysis of a database to which the search queries are directed.

9. The method as in claim 1, wherein analyzing search query submissions of a plurality of search engine users comprises collectively analyzing query submission histories of said plurality of users in an off-line processing mode.

10. The method as in claim 1, wherein analyzing search query submissions of a plurality of search engine users comprises generating, for each of a plurality of pairs of search terms, a score that reflects an occurrence count of search query submissions containing the respective pair of search terms.

11. The method as in claim 1, further comprising programmatically analyzing a search query submission that includes a first search term and a second search term, at least in part, by accessing the table to look up at least one additional term that is related to the first search term, and comparing a spelling of the additional term to the second search term.

12. The method as in claim 11, wherein programmatically analyzing the search query submission further comprises outputting a suggestion of the additional term to a user who made the search query submission in response to determining that the additional term and the second search term satisfy a spelling similarity test.

13. The method as in claim 1, wherein the table comprises a B-tree data structure.

14. The method as in claim 1, wherein analyzing the search query submissions comprises, by execution of code by said computer system, processing a log of search query submissions submitted to a search engine.

15. A system for identifying related terms, comprising:

a computer system that includes one or more computers, said computer system programmed to perform a method that comprises;

analyzing search query submissions of a plurality of search engine users to measure degrees to which particular search terms are related to each other, said degrees of relationship measured based at least partly on how frequently particular search terms co-occur in multiple-term search queries of users, wherein analyzing the search query submissions comprises weighting some search query submissions differently than other search query submissions such that some search query submissions are given different non-zero amounts of weight than other search query submissions in measuring said degrees to which particular search terms are related; and

generating a data set that reflects results of the analysis of the search query submissions of the plurality of users, said data set identifying specific search terms that are related to each other.

16. The system of claim 15, wherein the computer system is programmed to apply, to a search query submission of a user, a weight that depends on whether the user viewed an associated query result item.

17. The system of claim 15, wherein the computer system is programmed to give greater weight, in measuring said degrees to which particular search terms are related, to search query submissions that resulted in a user transaction than to search query submissions that did not result in a user transaction.

18. The system of claim 15, wherein the computer system is additionally programmed to use the data set to correct spelling errors in search query submissions of users.

19. The system of claim 15, wherein the computer system is additionally programmed to use the data set to select a term to use in place of a non-matching search term in a multi-term search query submitted by a user.

20. The system of claim 15, wherein the computer system is additionally programmed to measure said degrees to which particular search terms are related by performing a term co-occurrence analysis of a data repository to which the search queries are directed.

21. The system of claim 15, wherein the computer system is programmed to analyze the search query submissions of the plurality of search engine users in an off-line processing mode.

22. The system of claim 15, wherein the computer system is programmed to generate, for each of a plurality of pairs of search terms, a score that reflects an occurrence count of search query submissions containing the respective pair of search terms.

23. The system of claim 15, wherein the computer system is additionally programmed to analyze a search query submission that includes a first search term and a second search term, at least in part, by (1) accessing the data set to look up at least one additional term that is related to the first search term, and (2) comparing a spelling of the additional term to the second search term.

24. The system of claim 15, wherein the computer system is additionally programmed to use the data set to modify search query submissions of users.

25. Non-transitory computer storage that stores computer code that directs a computer system to perform a method that comprises:

analyzing search query submissions of a plurality of search engine users to measure degrees to which particular search terms are related to each other, said degrees of relationship measured based at least partly on how frequently particular search terms co-occur in multiple-term search queries of users, wherein analyzing the search query submissions comprises weighting some search query submissions differently than other search query submissions such that some search query submissions are given different non-zero amounts of weight than other search query submissions in measuring said degrees to which particular search terms are related; and

generating a data set that reflects results of the analysis of the search query submissions of the plurality of users, said data set identifying specific search terms that are related to each other.