System and method for optimal adaptive matching of users to most relevant entity and information in real-time

• US 6,134,532 A
• Filed: 11/14/1997
• Issued: 10/17/2000
• Est. Priority Date: 11/14/1997
• Status: Expired due to Term

• Current Assignee: Tumbleweed Holdings LLC (Equitable IP Corporation)
• Original Assignee: Aptex Software, Inc.
• Inventors: Lazarus, Michael A., Caid, William R., Pugh, Richard S., Kindig, Bradley D., Russell, Gerald S., Brown, Kenneth B., Dunning, Ted E., Carleton, Joel L.
• Primary Examiner(s): Voeltz, Emanuel Todd
• Assistant Examiner(s): Morgan, George D.
• Application Number:
  US08/971,091
• Time in Patent Office: 1,068 Days
• Field of Search: 705/14, 705/10, 705/1, 705/27, 705/20, 705/26, 707/532
• US Class Current: 705/14.25
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0224 based on user history
• CPC Class Codes: G06Q 30/0256 User search
• CPC Class Codes: G06Q 30/0264 based upon schedule
• CPC Class Codes: G06Q 30/0269 based on user profile or attribute

Abstract

A system and method for selecting and presenting personally targeted entities such as advertising, coupons, products and information content, based on tracking observed behavior on a user-by-user basis and utilizing an adaptive vector space representation for both information and behavior. The system matches users to entities in a manner that improves with increased operation and observation of user behavior. User behavior and entities (ads, coupons, products) and information (text) are all represented as content vectors in a unified vector space. The system is based on an information representation called content vectors that utilizes a constrained self organization learning technique to learn the relationships between symbols (typically words in unstructured text). Users and entities are each represented as content vectors.

Claims

1. A computerized system for associating an observed behavior with items, comprising:

a converter capable of converting the observed behavior to a behavior vector;

a profile adapter capable of modifying a profile vector with the behavior vector; and

a comparater capable of comparing the modified profile vector to a plurality of entity vectors, the entity vectors indicative of the items, so as to identify at least one entity vector closely associated with the observed behavior.

2. The computerized system as defined in claim 1, wherein the observed behavior is selected from the group consisting of:

submitting a query to a web site, requesting a web page, purchasing a product, visiting a merchant, visiting a web site, inquiring about a product, watching a TV show, and watching a movie.

3. The computerized system as defined in claim 1, wherein at least one of the items is selected from the group consisting of:

a coupon, an advertisement, a solicitation, information relating to a product, information relating to a set of services, a page, section or chapter of a book, a document, a newspaper article, a movie, a TV show, a web site, and a textual material.

4. The computerized system as defined in claim 1, wherein the computerized system is connected to a network, the network selected from the group consisting of:

an intranet, a local area network, and the Internet.

5. The computerized system as defined in claim 1, wherein the converter includes a page content vector lookup module which identifies the behavior vector based upon a page identifier.

6. The computerized system as defined in claim 1, wherein the observed behavior comprises a user query, and wherein the converter includes an entity content vector module for transforming the user query into behavior vectors based upon the component words of the user query.

7. The computerized system as defined in claim 1, wherein the profile adapter modifies the profile vector based upon at least one parameter selected from the group consisting of:

a threshold by which a behavior vector will be used instead of the entity vector, a learning rate for a profile update, a leaning rate for an entity update, an update rate for a query universe estimate of a mean, an update for a profile universe update of a mean, a forgetting factor for a set of profile vectors, a forgetting factor for a set of entity vectors, a mean of a set of entity vectors, a mean of a set profile vectors, and a mean of a set of behavior vectors.

8. The computerized system as defined in claim 1, wherein the comparater includes a vector closeness determination module to calculate the distance between the modified profile vector and any one of the entity vectors.

9. A system for selecting advertisements in a computer environment, comprising:

a database of electronic advertisements; and

an electronic advertisement management system, comprising;

a converter capable of converting an observed behavior of a user computing device in the computer environment to a behavior vector,

a profile adapter capable of modifying a profile vector indicative of the user with the behavior vector,

a comparater capable of comparing the modified profile vector to a plurality of entity vectors, the entity vectors indicative of the electronic advertisements, so as to identify at least one entity vector closely associated with the observed behavior, and

a selector accessing the electronic database with the identified entity vector so as to select at least one electronic advertisement to communicate to the user computing device.

10. The system as defined in claim 9, wherein the selector includes inventory management to allow selection of an entity that is under-selected according to the selection schedule and to inhibit the selection of an entity that is over-selected according to the selection schedule.

11. The system as defined in claim 9, wherein the selector includes inventory management to allow selection of the entity vector based upon a presentation delivery schedule.

12. The system as defined in claim 9, wherein the observed behavior is selected from the group consisting of:

submitting a query to a web site, requesting a web page, purchasing a product, visiting a merchant, visiting a web site, inquiring about a product, watching a TV show, and watching a movie.

13. The system as defined in claim 9, wherein the computer environment is connected to a network, the network selected from the group consisting of:

an intranet, a local area network, and the Internet.

14. The system as defined in claim 9, wherein the converter includes a page content vector lookup module which identifies the behavior vector based upon a page identifier.

15. The system as defined in claim 9, wherein the observed behavior is selected from the group consisting of:

a user query, a page view, or a purchase of a product.

16. The system as defined in claim 9, wherein the profile adapter modifies the profile vector based upon at least one parameter selected from the group consisting of:

a threshold by which a behavior vector will be used instead of the entity vector, a learning rate for a profile update, a leaning rate for an entity update, an update rate for a query universe estimate of a mean, an update for a profile universe update of a mean, a forgetting factor for a set of profile vectors, a forgetting factor for a set of entity vectors, a mean of a set of entity vectors, a mean of a set profile vectors, and a mean of a set of behavior vectors.

17. The system as defined in claim 9, wherein the comparater includes a vector closeness determination module to calculate the distance between the modified profile vector and any one of the entity vectors.

18. A computerized system for adapting an entity vector, comprising:

a converter capable of converting an observed behavior of a user into a behavior vector;

a profile adapter capable of modifying a profile vector indicative of the user based on the behavior vector; and

an entity adapter capable of modifying an entity vector indicative of an item based on the profile vector or the behavior vector.

19. The system as defined in claim 18, wherein at least one of the items is selected from the group consisting of:

a coupon, an advertisement, a solicitation, information relating to a product, information relating to a set of services, a page, a section or a chapter of a book, a document, a newspaper article, a movie, a TV show, a web site, and a textual material.

20. The system as defined in claim 18, wherein the observed behavior is selected from the group consisting of:

submitting a query to a web site, requesting a web page, purchasing, a product, visiting a merchant, visiting a web site, inquiring about a product, watching a TV show, and watching a movie.

21. The system as defined in claim 18, wherein the computerized system is connected to a network, the network selected from the group consisting of:

an intranet, a local area network, and the Internet.

22. The system as defined in claim 18, wherein the converter includes a page content vector lookup module which identifies the behavior vector based upon a page identifier.

23. The system as defined in claim 18, wherein the observed behavior is selected from the group consisting of:

a user query, a page view, or a purchase of a product.

24. The system as defined in claim 18, wherein the profile adapter modifies the profile vector based upon at least one parameter selected from the group consisting of:

a threshold by which a behavior vector will be used instead of the entity vector, a learning rate for a profile update, a learning rate for an entity update, an update rate for a query universe estimate of a mean, an update for a profile universe update of a mean, a forgetting factor for a set of profile vectors, a forgetting factor for a set of entity vectors, a mean of a set of entity vectors, a mean of a set profile vectors, and a mean of a set of behavior vectors.

25. The system as defined in claim 18, wherein the entity adapter modifies the entity vector based upon at least one parameter selected from the group consisting of:

a threshold by which a behavior vector will be used instead of the profile vector, a leaning rate for a profile update, a learning rate for an entity update, an update rate for a query universe estimate of a mean, an update for a profile universe update of a mean, a forgetting factor for a set of profile vectors, a forgetting factor for a set of entity vectors, a mean of a set of entity vectors, a mean of a set profile vectors, and a mean of a set of behavior vectors.

26. A system for generating a profile vector in a computer environment, comprising:

a converter capable of converting a plurality of observed behaviors of a user into an associated plurality of behavior vectors; and

a profile adapter capable of repeatedly modifying a profile vector indicative of the user based on the plurality of behavior vectors.

27. The system as defined in claim 26, wherein at least one of the plurality of observed user behaviors is selected from the group consisting of:

submitting a query to a web site, requesting a web page, purchasing a product, visiting a merchant, visiting a web site, inquiring about a product, watching a TV show, and watching a movie.

28. The system as defined in claim 26, wherein the computer environment is connected to a network, the network selected from the group consisting of:

an intranet, a local area network, and the Internet.

29. The system as defined in claim 26, wherein the observed behavior is selected from the group, consisting of:

a user query, a page view, or a purchase of a product.

30. The system as defined in claim 26, wherein the profile adapter includes modifies the profile vector based upon at least one parameter selected from the group consisting of:

a threshold by which a behavior vector will be used instead of the profile vector, a learning rate for a profile update, a learning rate for an entity update, an update rate for a query universe estimate of a mean, an update for a profile universe update of a mean, a forgetting factor for a set of profile vectors, a forgetting factor for a set of entity vectors, a mean of a set of entity vectors, a mean of a set profile vectors, and a mean of a set of behavior vectors.

31. A computerized system for adapting an entity vector, comprising:

a converter capable of converting an observed behavior of a user into a behavior vector; and

an entity adapter capable of modifying an entity vector indicative of an item based on the behavior vector.

32. The system as defined in claim 31, wherein observed behavior is selected from the group consisting of:

submitting a query to a web site, requesting a web page, purchasing a product, visiting a merchant, visiting a web site, inquiring about a product, watching a TV show, and watching a movie.

33. The system as defined in claim 31, wherein the computerized system is connected to a network, the network selected from the group consisting of:

an intranet, a local area network, and the Internet.

34. The system as defined in claim 31, wherein the converter includes a page content vector lookup module which identifies the behavior entity vector based upon a page identifier.

35. The system as defined in claim 31, wherein the observed behavior is selected from the group consisting of:

a user query, a page view, or a purchase of a product.

36. The system as defined in claim 31, wherein the item is selected from the group consisting of:

a coupon, an advertisement, a solicitation, information relating to a product, information relating to a set of services, a page, a section or a chapter of a book, a document, a newspaper article, a movie, a TV show, a web site, and a textual material.

37. A system for selecting advertisements in a computer environment, comprising:

a database of electronic advertisements; and

an electronic advertisement management system, comprising;

a converter capable of converting an observed behavior of a user computing device in the computer environment to a behavior vector,

a comparater capable of comparing the behavior vector to a plurality of entity vectors, the entity vectors indicative of the electronic advertisements, so as to identify at least one entity vector closely associated with the observed behavior, and

a selector accessing the electronic database with the identified entity vector so as to select at least one electronic advertisement to communicate to the user computing device.

38. The system as defined in claim 37, wherein the computer environment is connected to a network, the network selected from the group consisting of:

an intranet, a local area network, and the Internet.

39. The system as defined in claim 37, wherein the converter includes a page content vector lookup module which identifies the behavior vector based upon a page identifier.

40. The system as defined in claim 37, wherein the observed behavior is selected from the group consisting of:

a user query, a page view, or a purchase of a product.

41. The system as defined in claim 37, wherein the comparater includes a vector closeness determination module to calculate the distance between the behavior vector and any one of the entity vectors.

42. The system as defined in claim 37, wherein the selector includes inventory management to allow selection of the entity vector of an entity being under-selected according to a selection schedule and inhibit the selection of an entity that is over-selected according to the selection schedule.

43. The system as defined in claim 37, wherein the selector includes inventory management to allow selection of the entity based upon a presentation delivery schedule.

44. A method of associating an observed behavior with items on a computer including a data storage, comprising:

converting an observed behavior to a behavior vector;

modifying a profile vector with the behavior vector, and comparing the modified profile vector to a plurality of entity vectors, the entity vectors indicative of the items, so as to identify at least one entity vector closely associated with the observed behavior.

45. The method as defined in claim 44, wherein the observed user behavior is selected from the group consisting of:

submitting a query to a web site, requesting a web page, purchasing a product, visiting a merchant, visiting a web site, inquiring about a product, watching a TV show, and watching a movie.

46. The method as defined in claim 44, wherein at least one of the items is selected from the group consisting of:

a coupon, an advertisement, a solicitation, information relating to a product, information relating to a set of services, a page, a section or a chapter of a book, a document, a newspaper article, a movie, a TV show, a web site, and a textual material.

47. The method as defined in claim 44, wherein the converting step identifies the behavior vector based upon a page identifier.

48. The method as defined in claim 44, wherein the observed user behavior comprises a user query, and the converting step transforms the user query into vectors based upon the component words of the user query.

49. The method as defined in claim 44, wherein the comparing step calculates the distance between the modified profile vector and any one of the entity vectors.

50. A method of selecting advertisements in a computer environment, comprising:

providing a database of electronic advertisements;

converting an observed behavior of a user computing device in the computer environment to a behavior vector, modifying a profile vector indicative of the user with the behavior vector;

comparing the modified profile vector to a plurality of entity vectors, the entity vectors indicative of the electronic advertisements, so as to identify at least one entity vector closely associated with the observed behavior;

accessing the electronic database with the identified entity vector; and

selecting at least one electronic advertisement to communicate to the user computing device.

51. The method as defined in claim 50, wherein the selecting includes selecting an electronic advertisement that is under-selected according to a selection schedule and to inhibit the selection of an electronic advertisement that is over-selected according to the selection schedule.

52. The method as defined in claim 50, wherein the selecting includes selecting the entity vector based upon a presentation delivery schedule.

53. The method as defined in claim 50, wherein the modifying includes modifying the profile vector based upon at least one parameter selected from the group consisting of:

a threshold by which a behavior vector will be used instead of the entity vector, a learning rate for a profile update, a learning rate for an entity update, an update rate for a query universe estimate of a mean, an update for a profile universe update of a mean, a forgetting factor for a set of profile vectors, a forgetting factor for a set of entity vectors, a mean of a set of entity vectors, a mean of a set profile vectors, and a mean of a set of query vectors.

54. The method as defined in claim 50, wherein the electronic advertisement is communicated to the user computing device via a network, the network selected from the group consisting of an intranet, a local area network, and the Internet.

55. The method as defined in claim 50, wherein the observed user behavior is selected from the group consisting of:

submitting a query to a web site, requesting a web page, purchasing a product, visiting a merchant, visiting a web site, inquiring about a product, watching a TV show, and watching a movie.

56. The method as defined in claim 50, wherein the converting includes identifying the behavior vector based upon a page identifier.

57. The method as defined in claim 50, wherein the observed user behavior comprises a user query, and the converting includes transforming the user query into behavior vectors based upon the component words of the user query.

58. A method for adapting an entity vector on a computer including a data storage, comprising:

converting an observed behavior of a user into a behavior vector;

modifying a profile vector indicative of the user based on the behavior vector; and

modifying an entity vector indicative of an item based on the profile vector or the behavior vector.

59. The method as defined in claim 58, wherein at least one of the items is selected from the group consisting of:

a coupon, an advertisement, information relating to a solicitation, information relating to a product, a set of services, a page, a section or a chapter of a book, a document, a newspaper article, a movie, a TV show, a web site, and a textual material.

60. The method as defined in claim 58, wherein the modifying the profile vector includes modifying the profile vector based upon at least one parameter selected from the group consisting of:

a threshold by which a behavior vector will be used instead of the entity vector, a learning rate for a profile update, a learning rate for an entity update, an update rate for a query universe estimate of a mean, an update for a profile universe update of a mean, a forgetting factor for a set of profile vectors, a forgetting factor for a set of entity vectors, a mean of a set of entity vectors, a mean of a set profile vectors, and a mean of a set of behavior vectors.

61. The method as defined in claim 58, wherein the modifying the entity vector includes modifying the profile vector based upon at least one parameter selected from the group consisting of:

a threshold by which a behavior vector will be used instead of the profile vector, a learning rate for a profile update, a learning rate for an entity update, an update rate for a query universe estimate of a mean, an update for a profile universe update of a mean, a forgetting factor for a set of profile vectors, a forgetting factor for a set of entity vectors, a mean of a set of entity vectors, a mean of a set profile vectors, and a mean of a set of behavior vectors.

62. The method as defined in claim 58, wherein the computer is connected to a network, the network selected from the group consisting of:

an intranet, a local area network, and the Internet.

63. The method as defined in claim 58, wherein the observed user behavior is selected from the group consisting of:

submitting a query to a web site, requesting a web page, purchasing a product, visiting a merchant, visiting a web site, inquiring about a product, watching a TV show, and watching a movie.

64. The method as defined in claim 58, wherein the converting includes identifying identifies the behavior vector based upon a page identifier.

65. The method as defined in claim 58, wherein the observed user behavior comprises a user query, and the converting includes transforming the user query into vectors based upon the component words of the user query.

66. A method of generating a profile vector in a computer environment on a computer including a data storage, comprising:

converting a plurality of observed behaviors of a user into an associated plurality of behavior vectors; and

repeatedly modifying a profile vector indicative of the user based on the plurality of behavior vectors.

67. The method as defined in claim 66, wherein the computer environment is connected to a network, the network selected from the group consisting of:

an intranet, a local area network, and the Internet.

68. The method as defined in claim 66, wherein the observed user behavior is selected from the group consisting of:

submitting a query to a web site, requesting a web page, purchasing a product, visiting a merchant, visiting a web site, inquiring about a product, watching a TV show, and watching a movie.

69. The method as defined in claim 66, wherein the converting includes identifying the behavior vector based upon a page identifier.

70. The method as defined in claim 66, wherein the observed user behavior comprises a user query, and the converting includes transforming the user query into vectors based upon the component words of the user query.

71. A method of adapting an entity vector on a computer including a data storage, comprising:

converting an observed behavior of a user into a behavior vector, and

modifying an entity vector indicative of an item based on the behavior vector.

72. The method as defined in claim 71, wherein at least one of the items are selected from the group consisting of:

a coupon, an advertisement, a solicitation, information relating to a product, information relating to a set of services, a page, a section or chapter of a book, a document, a newspaper article, a movie, a TV show, a web site, and a textual material.

73. The method as defined in claim 71, wherein the computer is connected to a network, the network selected from the group consisting of:

an intranet, a local area network, and the Internet.

74. The method as defined in claim 71, wherein the observed user behavior is selected from the group consisting of:

submitting a query to a web site, requesting a web page, purchasing a product, visiting a merchant, visiting a web site, inquiring about a product, watching a TV show, and watching a movie.

75. The method as defined in claim 71, wherein the converting includes identifying the behavior vector based upon a page identifier.

76. The method as defined in claim 71, wherein the observed user behavior comprises a user query, and the converting includes transforming the user query into vectors based upon the component words of the user query.

77. The method as defined in claim 71, wherein the modifying step including modifying for the entity vector modifies the behavior vector based upon at least one parameter selected from the group consisting of:

a threshold by which a behavior vector will be used instead of the profile vector, a learning rate for a profile update, a learning rate for an entity update, an update rate for a query universe estimate of a mean, an update for a profile universe update of a mean, a forgetting factor for a set of profile vectors, a forgetting factor for a set of entity vectors, a mean of a set of entity vectors, a mean of a set profile vectors, and a mean of a set of behavior vectors.

78. A method of selecting advertisements in a computer including a data storage, comprising:

providing a database of electronic advertisements;

converting an observed behavior of a user computing device in the computer to a behavior vector;

comparing the behavior vector to a plurality of entity vectors, the entity vectors indicative of the electronic advertisements, so as to identify at least one entity vector closely associated with the observed behavior;

accessing the electronic database with the identified entity vector, and

selecting at least one electronic advertisement to communicate to the user computing device.

79. The method as defined in claim 78, wherein the selecting includes selecting an electronic advertisement that is under-selected according to a selection schedule and inhibiting the selection of an electronic advertisement that is over-selected according to the selection schedule.

80. The method as defined in claim 78, wherein the selecting includes selecting the entity vector based upon a presentation delivery schedule.

81. The method as defined in claim 78, wherein the comparing includes calculating the distance between the behavior vectors to any one of the entity vectors.

82. The method as defined in claim 78, wherein the at least one electronic advertisement is communicated to the user computing device via a network, the network selected from the group consisting of:

an intranet, a local area network, and the Internet.

83. The method as defined in claim 78, wherein the observed user behavior is selected from the group consisting of:

submitting a query to a web site, requesting a web page, purchasing a product, visiting a merchant, visiting a web site, inquiring about a product, watching a TV show, and watching a movie.

84. The method as defined in claim 78, wherein the converting includes identifying the behavior vector based upon a page identifier.

85. The method as defined in claim 78, wherein the observed user behavior comprises a user query, and the converting includes transforming the user query into vectors based upon the component words of the user query.