Method for detecting and fulfilling an information need corresponding to simple queries

• US 20060259510A1
• Filed: 07/21/2006
• Published: 11/16/2006
• Est. Priority Date: 04/26/2000
• Status: Abandoned Application

• Current Assignee: SAS Institute Incorporated
• Original Assignee: SAS Institute Incorporated
• Inventors: Schabes, Yves, Roche, Emmanuel
• Application Number:
  US11/490,719
• Publication Number:
  US 20060259510A1
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/3325 Reformulation based on results of preceding query
• CPC Class Codes: G06F 16/3329 Natural language query formulation or dialogue systems
• CPC Class Codes: G06F 16/3331 Query processing
• CPC Class Codes: G06F 16/3332 Query translation
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 16/9532 Query formulation
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching
• CPC Class Codes: Y10S 707/99936 Pattern matching access

Abstract

Described are techniques used in fulfilling an information need not formulated as a natural language question. A query is accepted that includes one or more words not formulated as a natural language question. The query is syntactically analyzed in order to extract a concept from the query. Documents are identified containing the one or more words. The analyzed query is matched to one or more question keys in a structures repository. A best match is determined between the analyzed query and a question key of the structures repository. The key of the at least one best key match is mapped to partially specified queries and the key of the best question key match is mapped to questions representing an information need corresponding to the query. The partially specified queries and questions are matched to potential answers contained in a collection of documents. One or more matching answers are returned based on a metric.

Claims

1. A method of fulfilling an information need comprising:

obtaining an input query including at least one user-specified search term;

converting said input query into a converted input query; and

using said converted input query to obtain a result wherein said result does not include said at least one user-specified search term included in said input query.

2. The method of claim 1, wherein said result is an answer, said input query is not in the form of a natural language question, and said converted input query is a natural language question.

3. The method of claim 2, further comprising:

using a question—

answering engine to obtain said answer; and

performing a document search using said answer as a search term.

4. The method of claim 3, further comprising:

performing a document search using said at least one term included in said input query.

5. The method of claim 1, wherein said converted input query is an affirmative statement including an unspecified portion, and the method further comprises:

searching documents for candidate results for the unspecified portion; and

selecting said result from said candidate results.

6. The method of claim 5, further comprising:

performing a document search using said result as a search term.

7. The method of claim 6, further comprising:

performing a document search using said at least one term included in said input query.

8. The method of claim 5, further comprising:

analyzing said input query to extract syntactic and contextual information producing an analyzed query;

determining a best key match for said analyzed query by matching said analyzed query with key terms associated with a structure;

forming at least one partially specified query using other data included in said structure;

searching documents using said at least one partially specified query to obtain said candidate results, said at least one partially specified query being said converted query; and

selecting said result from said candidate results.

9. The method of claim 2, further comprising:

analyzing said input query to extract syntactic and contextual information producing an analyzed query;

determining a best key match for said analyzed query by matching said analyzed query with key terms associated with a structure;

forming at least one question using other data included in said structure, said at least one question being said converted query; and

using a question-answering engine to obtain said answer to said at least one question.

10. The method of claim 8, wherein said other data includes at least one statement format.

11. The method of claim 10, wherein said other data includes synonyms for said at least one statement format.

12. The method of claim 11, further comprising:

forming said at least one partially specified query using said at least one statement format;

determining whether candidate results are obtained using said at least one statement format; and

conditionally forming and using other partially specified queries using said synonyms if said at least one statement format did not return candidate results.

13. The method of claim 8, further comprising:

formulating partially specified queries using various fields of other data specified in said structure in accordance with a processing order determined by said structure.

14. The method of claim 1, further comprising:

analyzing said input query to extract syntactic and contextual information producing an analyzed query;

extracting information from said analyzed input query to determine if said input query is asking a question; and

determining said result using one of said question and an affirmative statement of said question.

15. A data structure stored in a computer memory comprising:

at least one key field describing an input query format;

at least one of;

a statement format field specifying an alternate form in which data included in said input query format may be specified as an affirmative statement; and

a question format field specifying data included in said input query format as a question.

16. The data structure of claim 15, further comprising:

at least one synonym statement format field specifying an alternate form in which data included in said statement format may be specified.

17. The data structure of claim 16, wherein said input query format includes at least one element that is a concept representative of more specific data.

18. The data structure of claim 17, wherein said input query format includes at least one element that matches a term from an input query.

19. The data structure of claim 15, further comprising:

a database statement format field that includes a database query statement including data related to an input query in accordance with said input query format.

20. The data structure of claim 15, further comprising:

an application statement format field that includes an application invocation statement including data related to an input query in accordance with said input query format.

21. A method for determining a result to an input query comprising:

determining whether said input query matches an input query format;

if said input query matches said specified input query format, determining at least one alternate form in which data included in said input query format may be specified as at least one of;

an affirmative statement of a question and a question; and

determining said result using said at least one of said affirmative statement and said question, said result not including user-specified search terms of said input query.

22. The method of claim 21, wherein said determining whether said input query matches a specified input query format uses a key field specified in a defined structure.

23. The method of claim 22, wherein said determining at least one alternate form uses at least one of:

a statement field included in said defined structure if specifying an affirmative statement, and a question field included in said defined structure if specifying a question.

24. The method of claim 21, further comprising:

searching documents using said at least one of said affirmative statement and said question.

25. The method of claim 21, wherein said at least one alternate form includes an affirmative statement, and the method further comprising:

specifying at least one synonym for said affirmative statement specifying an alternate form in which data included in said statement format may be specified.

26. The method of claim 21, wherein said input query format includes at least one element that is a concept representative of more specific data.

27. The method of claim 21, wherein said input query format includes at least one element that matches a term from an input query.

28. The method of claim 21, further comprising:

performing a database search using a database statement format field of a predefined structure that includes a database query statement including data related to an input query in accordance with said input query format.

29. The method of claim 21, further comprising:

invoking an application using an application statement format field that includes data related to an input query in accordance with said input query format.

30. The method of claim 23, wherein said at least one alternate form is a plurality of alternate forms, each of said alternate forms determined using a field from said defined structure in an order in accordance with an ordering of fields within said defined structure.

31. The method of claim 21, wherein said input query includes specific words, said specified input query includes a concept related to at least one of said specific words arranged in a context sensitive order, the method further comprising:

substituting said specific word for said concept in said at least one alternate form; and

performing at least one of;

searching at least one document, using a question-answer engine, and searching a database using said alternate form to determine said result.

32. The method of claim 31, wherein a defined structure includes fields specifying statements utilized in searching at least one document, invoking a question-answer engine, and searching a database, said defined structure specifying an order in which said statements are utilized.

33. A computer program product for fulfilling an information need comprising:

executable code that obtains an input query including at least one user-specified search term;

executable code that converts said input query into a converted input query; and

executable code that uses said converted input query to obtain a result wherein aid result does not include said at least one user-specified search term included in said input query.

34. The computer program product of claim 33, wherein said result is an answer, said input query is not in the form of a natural language question, and said converted input query is a natural language question.

35. The computer program product of claim 34, further comprising:

executable code that uses a question—

answering engine to obtain said answer; and

executable code that performs a document search using said answer as a search term.

36. The computer program product of claim 35, further comprising:

executable code that performs a document search using said at least one term included in said input query.

37. The computer program product of claim 33, wherein said converted input query is an affirmative statement including an unspecified portion, and the computer program product further comprises:

executable code that searches documents for candidate results for the unspecified portion; and

executable code that selects said result from said candidate results.

38. The computer program product of claim 37, further comprising:

executable code that performs a document search using said result as a search term.

40. The computer program product of claim 37, further comprising:

executable code that analyzes said input query to extract syntactic and contextual information producing an analyzed query;

executable code that determines a best key match for said analyzed query by matching said analyzed query with key terms associated with a structure;

executable code that forms at least one partially specified query using other data included in said structure;

executable code that searches documents using said at least one partially specified query to obtain said candidate results, said at least one partially specified query being said converted query; and

executable code that selects said result from said candidate results.

41. The computer program product of claim 34, further comprising:

executable code that analyzes said input query to extract syntactic and contextual information producing an analyzed query;

executable code that determines a best key match for said analyzed query by matching said analyzed query with key terms associated with a structure;

executable code that forms at least one question using other data included in said structure, said at least one question being said converted query; and

executable code that uses a question-answering engine to obtain said answer to said at least one question.

42. The computer program product of claim 40, wherein said other data includes at least one statement format.

43. The computer program product of claim 42, wherein said other data includes synonyms for said at least one statement format.

44. The computer program product of claim 43, further comprising:

executable code that forms said at least one partially specified query using said at least one statement format;

executable code that determines whether candidate results are obtained using said at least one statement format; and

executable code that conditionally forms and uses other partially specified queries using said synonyms if said at least one statement format does not return candidate results.

45. The computer program product of claim 40, further comprising:

executable code that formulates partially specified queries using various fields of other data specified in said structure in accordance with a processing order determined by said structure.

46. The computer program product of claim 33, further comprising:

executable code that analyzes said input query to extract syntactic and contextual information producing an analyzed query;

executable code that extracts information from said analyzed input query to determine if said input query is asking a question; and

executable code that determines said result using one of said question and an affirmative statement of said question.

39. The computer program product of claim 39, further comprising:

executable code that performs a document search using said at least one term included in said input query.

47. A computer program product for determining a result to an input query comprising:

executable code that determines whether said input query matches an input query format;

executable code that, if said input query matches said specified input query format, determines at least one alternate form in which data included in said input query format may be specified as at least one of;

an affirmative statement of a question and a question; and

executable code that determines said result using said at least one of said affirmative statement and said question, said result not including user-specified search terms of said input query.

48. The computer program product of claim 47, wherein said executable code that determines whether said input query matches a specified input query format includes machine executable code that uses a key field specified in a defined structure.

49. The computer program product of claim 48, wherein said executable code that determines at least one alternate form includes executable code that uses at least one of:

a statement field included in said defined structure if specifying an affirmative statement, and a question field included in said defined structure if specifying a question.

50. The computer program product of claim 47, further comprising:

executable code that searches documents using said at least one of said affirmative statement and said question.

51. The computer program product of claim 47, wherein said at least one alternate form includes an affirmative statement, and the computer program product further comprising:

executable code that specifies at least one synonym for said affirmative statement specifying an alternate form in which data included in said statement format may be specified.

52. The computer program product of claim 47, wherein said input query format includes at least one element that is a concept representative of more specific data.

53. The computer program product of claim 47, wherein said input query format includes at least one element that matches a term from an input query.

54. The computer program product of claim 47, further comprising:

executable code that performs a database search using a database statement format field of a predefined structure that includes a database query statement including data related to an input query in accordance with said input query format.

55. The computer program product of claim 47, further comprising:

executable code that invokes an application using an application statement format field that includes data related to an input query in accordance with said input query format.

56. The computer program product of claim 49, wherein said at least one alternate form is a plurality of alternate forms, each of said alternate forms determined using a field from said defined structure in an order in accordance with an ordering of fields within said defined structure.

57. The computer program product of claim 47, wherein said input query includes specific words, said specified input query includes a concept related to at least one of said specific words arranged in a context sensitive order, the computer program product further comprising:

executable code that substitutes said specific word for said concept in said at least one alternate form; and

executable code that performs at least one of;

searching at least one document, using a question-answer engine, and searching a database using said alternate form to determine said result.

58. The computer program product of claim 57, wherein a defined structure includes fields specifying statements utilized in searching at least one document, invoking a question-answer engine, and searching a database, said defined structure specifying an order in which said statements are utilized.

59. A method of fulfilling an information need not formulated as a natural language question, comprising:

accepting a query that includes one or more words not formulated as a natural language question but which corresponds to an information need; and

returning one or more answers satisfying the information need.

60. The method of claim 59, further comprising:

syntactically analyzing the query in order to extract a concept from the query;

matching the analyzed query to one or more keys in a structures repository;

determining a best match between the analyzed query and a key of the structures repository;

mapping the key of the best match to partially specified queries representing an information need corresponding to the query; and

matching the partially specified queries to potential answers contained in a collection of documents.

62. The method of claim 59, further comprising:

returning document identifiers corresponding to the documents containing the matching answers.

61. A method of fulfilling an information need not formulated as a natural language question, comprising:

accepting a query including one or more words not formulated as a natural language question;

syntactically analyzing the query in order to extract a concept from the query;

identifying and returning documents containing the one or more words;

matching the analyzed query to one or more keys in a structures repository;

determining a best match between the analyzed query and a key of the structures repository;

mapping the key of the best match to partially specified queries representing an information need corresponding to the query;

matching the partially specified queries to potential answers contained in a collection of documents; and

returning one or more matching answers based on a metric.

63. A method of fulfilling an information need not formulated as a natural language question, comprising:

accepting a query including one or more words not formulated as a natural language question;

identifying and returning documents containing the one or more words; and

returning one or more answers satisfying an information need corresponding to the query.

64. A method of fulfilling an information need not formulated as a natural language question, comprising:

accepting a query including one or more words not formulated as a natural language question;

syntactically analyzing the query in order to extract a concept from the query;

identifying and returning documents containing the one or more words;

matching the analyzed query to one or more keys in a structures repository;

determining a best match between the analyzed query and a key of the structures repository;

mapping the key of the best match to partially specified queries representing an information need corresponding to the query;

matching the partially specified queries to potential answers contained in a collection of documents; and

returning one or more matching answers based on a metric.

65. A method of fulfilling an information need not formulated as a natural language question, comprising:

accepting a query including one or more words not formulated as a natural language question;

transforming the query into one or more natural language questions; and

processing the natural language questions to obtain answers from a collection of documents.

66. The method of claim 65, wherein the transforming further comprises:

syntactically analyzing the query in order to extract a concept from the query;

matching the analyzed query to one or more question keys in a structures repository;

determining a best match between the analyzed query and a question key of the structures repository; and

mapping the question key of the best match to natural language questions representing an information need corresponding to the query.

67. The method of claim 66, wherein processing the natural language questions further comprises:

matching the natural language questions to potential answers contained in a collection of documents; and

returning one or more matching answers based on a metric.

68. A method of fulfilling an information need not formulated as a natural language question, comprising:

accepting a query comprised of one or more words not formulated as a natural language question;

syntactically analyzing the query in order to extract a concept from the query;

identifying and returning documents containing the one or more words;

matching the analyzed query to one or more question keys in a structures repository;

determining a best match between the analyzed query and a question key of the structures repository;

mapping the question key of the best match to questions representing an information need corresponding to the query;

matching the questions to potential answers contained in a collection of documents; and

returning one or more matching answers based on a metric.

69. A method of fulfilling an information need not formulated as a natural language question, comprising:

accepting a query including one or more words not formulated as a natural language question;

identifying and returning documents containing the one or more words;

transforming the query into one or more natural language questions; and

processing the natural language questions to obtain answers from a collection of documents.

70. The method of claim 69, wherein the transforming further comprises:

syntactically analyzing the query in order to extract a concept from the query;

matching the analyzed query to one or more question keys in a structures repository;

determining a best match between the analyzed query and a question key of the structures repository; and

mapping the question key of the best match to natural language questions representing an information need corresponding to the query.

71. The method of claim 70, wherein processing the natural language questions further comprises:

matching the natural language questions to potential answers contained in a collection of documents; and

returning one or more matching answers based on a metric.

72. A method of fulfilling an information need not formulated as a natural language question, comprising:

accepting a query comprised of one or more words not formulated as a natural language question;

syntactically analyzing the query in order to extract a concept from the query;

identifying documents containing the one or more words;

matching the analyzed query to one or more question keys in a structures repository;

determining a best match between the analyzed query and a question key of the structures repository;

mapping the question key of the best match to questions representing an information need corresponding to the query;

matching the questions to potential answers contained in a collection of documents; and

returning one or more matching answers based on a metric.

73. A method of fulfilling an information need not formulated as a natural language question, comprising:

accepting a query including at least one word not formulated as a natural language question;

syntactically analyzing the query producing an analyzed query in order to extract a concept from the query;

matching the analyzed query to at least one of a key and a question key in a structures repository;

determining at least one best match between the analyzed query and said at least one of a key and question key of the structures repository;

mapping the key of the at least one best key match to partially specified queries and the key of the best question key match to questions representing an information need corresponding to the query;

matching the partially specified queries and questions to potential answers contained in a collection of documents; and

returning one or more matching answers based on a metric.