Systems and methods for ranking implicit search results

• US 20070276829A1
• Filed: 03/31/2004
• Published: 11/29/2007
• Est. Priority Date: 03/31/2004
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Wang, Niniane, Lawrence, Stephen
• Granted Patent: US 7,693,825 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/3334 Selection or weighting of terms from queries, including natural language queries
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques

Abstract

Systems and methods for ranking implicit search queries are described. In one embodiment a method comprising receiving an event, the event comprising user interaction with an article on a client device, wherein the article is capable of being associated with at least one of a plurality of client applications, extracting at least one keyword from the event, generating a query based at least in part on the at least one keyword, performing a search based at least in part on the query to determine a result set, wherein the result set comprises one or more article identifiers associated with articles comprising the at least one keyword, and determining a ranking for each of the one or more article identifiers comprising the result set is described.

Claims

1. A method of ranking article identifiers of a result set from an implicit query implied from a user'"'"'s current context, the method comprising:

receiving an event concerning the user'"'"'s current context, wherein the event comprises a user interaction with an article stored on a local client device, wherein the article is associated with at least one of a plurality of client applications;

extracting at least one keyword from the event;

generating an implicit query based at least in part on the at least one keyword;

performing a search of events based at least in part on the implicit query to determine a result set, wherein the result set comprises one or more article identifiers associated with articles relevant to the implicit query; and

ranking the article identifiers.

2. The method of claim 1, wherein ranking the article identifiers is based at least in part on a preference of a current user.

3. The method of claim 2, wherein the preference of the current user is based at least in part on click-through data.

4. The method of claim 2, wherein the preference of the current user is based at least in part on file type.

5. The method of claim 1, wherein ranking the article identifiers is based at least in part on meta-data associated with an article.

6. The method of claim 5, wherein the meta-data comprises at least one of bolding, highlighting, italicizing, font color, or heading data.

7. The method of claim 1, wherein ranking the article identifiers is based at least in part on a term frequency and a document frequency.

8. The method of claim 7, wherein ranking the article identifiers comprises determining a rank that is proportional to the log of a sum of a first constant plus the term frequency and inversely proportional to the log of a sum of a second constant plus the document frequency.

9. The method of claim 1, wherein ranking the article identifiers comprises determining a rank that is proportional to the log of a sum of a constant plus a term frequency and inversely proportional to an output of a mapping function that maps ranges of document frequency into constants.

10. The method of claim 1, wherein ranking the article identifiers is based at least in part on number data.

11. The method of claim 10, wherein the number data comprises a number of letters in the keyword.

12. The method of claim 10, wherein the number data comprises whether a keyword comprises numbers.

13. The method of claim 1, wherein ranking the article identifiers is based at least in part on capitalization data.

14. The method of claim 1, wherein ranking the article identifiers is based at least in part on a number of sources from which the keyword was extracted.

15. The method of claim 1, wherein ranking the article identifiers is based at least in part on a number of result sets in which the result appears.

16. The method of claim 1, wherein the keywords are associated with keyword ranking scores.

17. The method of claim 16, wherein ranking the article identifiers is based at least in part on the keyword ranking scores.

18. The method of claim 17, wherein ranking the article identifiers comprises assigning a higher ranking to article identifiers associated with articles containing higher ranked keywords.

19. The method of claim 1, wherein extracting at least one keyword from an event comprises extracting a keyword from at least one of recently typed words, an entire document, a selected portion of a document, or words surrounding a cursor.

20. The method of claim 1, wherein extracting at least one keyword from an event comprises determining proper names.

21. The method of claim 20, wherein determining proper names comprises crawling at least one article.

54. The method of claim 1, wherein the article is a document on the client device, and wherein the event comprises an addition of words to the document.

55. The method of claim 1, wherein the article is a document on the client device, and wherein the event comprises a placement of a cursor near words in the document.

56. The method of claim 1, wherein the article is associated with one client application selected from a group consisting of a word processing program, a spreadsheet program, a presentation program, an e-mail program, an instant messenger program, and a database program.

22. A method of outputting article identifiers of a result set from an implicit query implied from a user'"'"'s current context, the method comprising:

receiving an event concerning the user'"'"'s current context, wherein the event comprises a user interaction with an article stored on a local client device, wherein the article is associated with at least one of a plurality of client applications;

extracting at least one keyword from the event;

generating an implicit query based at least in part on the at least one keyword;

performing a search based at least in part on the implicit query to determine a result set, wherein the result set comprises one or more article identifiers associated with articles comprising the at least one keyword;

filtering the result set based on a threshold; and

outputting the article identifiers associated with the filtered result set.

23. The method of claim 22, wherein the threshold comprises a number of keywords.

24. The method of claim 22, wherein the threshold comprises a minimum weighting score based at least in part on one or more of a number of keywords multiplier, a source multiplier, and a time multiplier.

25. The method of claim 22, further comprising determining a ranking score for each of the one or more articles identifiers.

26. The method of claim 25, further comprising arranging the article identifiers based at least in part on ranking score.

27. A computer-readable medium containing program code for ranking article identifiers of a result set from an implicit query implied from a user'"'"'s current context, the program code comprising:

program code for receiving an event concerning the user'"'"'s current context, wherein the event comprises a user interaction with an article stored on a local client device, wherein the article is associated with at least one of a plurality of client applications;

program code for extracting at least one keyword from the event;

program code for generating an implicit query based at least in part on the at least one keyword;

program code for performing a search of events based at least in part on the implicit query to determine a result set, wherein the result set comprises one or more article identifiers associated with articles relevant to the implicit query; and

program code for ranking the article identifiers.

28. The computer-readable medium of claim 27, wherein ranking the article identifiers is based at least in part on a preference of a current user.

29. The computer-readable medium of claim 28, wherein the preference of the current user is based at least in part on click-through data.

30. The computer-readable medium of claim 28, wherein the preference of the current user is based at least in part on file type.

31. The computer-readable medium of claim 27, wherein ranking the article identifiers is based at least in part on meta-data associated with an article.

32. The computer-readable medium of claim 31, wherein the meta-data comprises at least one of bolding, highlighting, italicizing, font color, or heading data.

33. The computer-readable medium of claim 27, wherein ranking the article identifiers is based at least in part on a term frequency and a document frequency.

34. The computer-readable medium of claim 33, wherein ranking the article identifiers comprises determining a rank that is proportional to the log of the sum of a first constant plus the term frequency and inversely proportional to the log of the sum of a second constant plus the document frequency.

35. The computer-readable medium of claim 27, wherein ranking the article identifiers comprises determining a rank that is proportional to the log of the sum of a constant plus a term frequency and inversely proportional to the output of a mapping function that maps ranges of document frequency into constants.

36. The computer-readable medium of claim 27, wherein ranking the article identifiers is based at least in part on number data.

37. The computer-readable medium of claim 36, wherein the number data comprises a number of letters in the keyword.

38. The computer-readable medium of claim 36, wherein the number data comprises whether a keyword comprises numbers.

39. The computer-readable medium of claim 27, wherein ranking the article identifiers is based at least in part on capitalization data.

40. The computer-readable medium of claim 27, wherein ranking the article identifiers is based at least in part on a number of sources from which the keyword was extracted.

41. The computer-readable medium of claim 27, wherein ranking the article identifiers is based at least in part on a number of result sets in which the result appears.

42. The computer-readable medium of claim 27, wherein the keywords are associated with keyword ranking scores.

43. The computer-readable medium of claim 42, wherein ranking the article identifiers is based at least in part on the keyword ranking scores.

44. The computer-readable medium of claim 43, wherein ranking the article identifiers comprises assigning a higher ranking to article identifiers associated with articles containing higher ranked keywords.

45. The computer-readable medium of claim 27, wherein extracting at least one keyword from an event comprises extracting a keyword from at least one of recently typed words, an entire document, a selected portion of a document, or words surrounding a cursor.

46. The computer-readable medium of claim 27, wherein extracting at least one keyword from an event comprises determining proper names.

47. The computer-readable medium of claim 46, wherein determining proper names comprises crawling at least one article.

48. A computer-readable medium containing program code for outputting article identifiers from a result set from an implicit query implied from a user'"'"'s current context, the program code comprising:

program code for receiving an event concerning the user'"'"'s current context, wherein the event comprises a user interaction with an article stored on a local client device, wherein the article is associated with at least one of a plurality of client applications;

program code for extracting at least one keyword from the event;

program code for generating an implicit query based at least in part on the at least one keyword;

program code for performing a search based at least in part on the implicit query to determine a result set, wherein the result set comprises one or more article identifiers associated with articles comprising the at least one keyword;

program code for filtering the result set based on a threshold; and

program code for outputting the article identifiers associated with the filtered result set.

49. The computer-readable medium of claim 48, wherein the threshold comprises a number of keywords.

50. The computer-readable medium of claim 48, wherein the threshold comprises a minimum weighting score based at least in part on one or more of a number of keywords multiplier, a source multiplier, and a time multiplier.

51. The computer-readable medium of claim 48, further comprising determining a ranking score for each of the one or more articles identifiers.

52. The computer-readable medium of claim 51, further comprising arranging the article identifiers based at least in part on ranking score.

53. A method of ranking article identifiers of a result set from an implicit query implied from a user'"'"'s current context, the method comprising:

receiving a contextual event concerning the user'"'"'s current context, the event comprising a user'"'"'s modification of a file stored on a local client device;

extracting at least one keyword from the contextual event;

generating an implicit query based at least in part on the at least one keyword;

performing a search based at least in part on the implicit query to determine a result set, wherein the result set comprises one or more article identifiers associated with articles comprising the at least one keyword;

determining a ranking score for the one or more article identifiers based on one or more of;

user preference data, click-through data, file type, meta-data, term frequency, inverse document frequency, number data, capitalization data, proper names, number of sources, and number of queries; and

ranking the one or more article identifiers in the result set based on the ranking score.