Rule-based management of adaptive models and agents

• US 8,099,376 B2
• Filed: 02/22/2008
• Issued: 01/17/2012
• Est. Priority Date: 02/22/2007
• Status: Active Grant

• Current Assignee: Fair Isaac Corporation
• Original Assignee: Fair Isaac Corporation
• Inventors: Serrano-Morales, Carlos A., Berlioz-Matignon, Carole-Ann, Mangin, Franck
• Primary Examiner(s): Gaffin, Jeffrey A
• Assistant Examiner(s): CHANG, LI WU
• Application Number:
  US12/036,161
• Publication Number:
  US 20080208786A1
• Time in Patent Office: 1,425 Days
• Field of Search: 706/47
• US Class Current: 706/47
• CPC Class Codes: G06N 5/025 Extracting rules from data

Abstract

A computer-implemented method described for making a recommendation with respect to a plurality of items using a plurality of adaptive models or agents. The described method includes receiving one or more business rules; receiving a request for a recommendation; receiving attributes relating to the request; activating one or more adaptive agents such that each activated adaptive agent generates one or more recommendations with respect to the items based at least in part on an evaluation of prior outcomes relating to the items; selecting from among the one or more recommendations at least one final recommendation; and displaying the at least one final recommendation to a user. Related apparatus, systems, techniques and articles are also described.

Claims

1. A system for making a recommendation with respect to a plurality of items using a plurality of adaptive agents comprising:

a plurality of adaptive agents each programmed to generate one or more recommendations with respect to one or more of the items based at least in part on an evaluation of prior outcomes relating to the items;

a management system arranged and programmed to receive one or more business rules, and manage the plurality of adaptive agents according to the business rules such that at least one final recommendation is selected from among the one or more recommendations from the plurality of adaptive agents; and

a user interface adapted to display at least one final recommendation to a user;

wherein at least some of adaptive agents have variable learning rates, the learning rate influencing a degree on which prior outcomes are relied upon when selecting a choice.

2. A system according to claim 1, wherein each of the plurality of adaptive agents considers a set of the items in generating the one or more recommendations, and wherein the management system defines the set of the items by selecting items from the plurality of items based at least in part on the received business rules.

3. A system according to claim 2, wherein the business rules include item eligibility rules which are used to select the set of items, the eligibility rules including one or more types selected from the group comprising:

geographical restrictions, customer-income restrictions, customer purchases, history restrictions, demographic restrictions, time of the day restrictions, and business condition-related restrictions.

4. A system according to claim 1, wherein the management system selects one or more of the plurality of adaptive agents according to the business rules such that the at least one final recommendation originates from one of the selected adaptive agents.

5. A system according to claim 4, wherein the management system receives a request for a recommendation and based on the business rules routes the request to a single adaptive agent.

6. A system according to claim 1, wherein the business rules include one or more business goals, and the management system selects one final recommendation based on the one or more business goals.

7. A system according to claim 6, wherein the one or more business goals is selected from the group comprising:

revenue maximization, profit maximization, customer satisfaction, customer retention, loss minimization, market penetration, and increase in market share.

8. A system according to claim 1, wherein the management system initiates new adaptive agents and retires existing adaptive agents based on the received business rules.

9. A system according to claim 8, wherein the business rules include rules having one or more parameters selected from the group comprising:

how often agents are created, how many agents should be instantiated concurrently, and how long an agent can remain active before retirement.

10. A system according to claim 1, wherein the management system receives and stores information on outcomes relating to items.

11. A system according to claim 10, wherein the user interface receives the information on outcomes relating to items and sends the information to the management system.

12. A system according to claim 10, wherein the management system determines whether the information has a correlation with one or more prior recommendations.

13. A system according to claim 1, further comprising a maintenance interface arranged and programmed to receive from maintenance users the business rules.

14. A system according to claim 13, wherein the maintenance interface is further arranged and programmed to receive new items for inclusion in the plurality of items.

15. A system according to claim 14, wherein the maintenance interface is further arranged and programmed to receive initial biases with respect to the new items.

16. A system according to claim 1, wherein the adaptive agents are adaptive decisioning agents.

17. A system according to claim 1, wherein the items include items having one or more types selected from the group comprising:

products, services, offers for products, offers for services, suppliers, and agencies.

18. A system according to claim 1, wherein the user interface is adapted to receive a request for a recommendation with respect to the plurality of items.

19. A system according to claim 18, wherein the request is received from the user.

20. A system according to claim 18, wherein the request is generated automatically by the user interface.

21. A computer-implemented method for making a recommendation with respect to a plurality of items using a plurality of adaptive agents comprising:

receiving one or more business rules;

receiving a request for a recommendation;

receiving attributes relating to the request;

activating one or more adaptive agents such that each activated adaptive agent generates one or more recommendations with respect to the one or more items based at least in part on an evaluation of prior outcomes relating to the items;

selecting from among the one or more recommendations at least one final recommendation; and

displaying the at least one final recommendation to a user;

wherein at least some of adaptive agents have variable learning rates, the learning rate influencing a degree on which prior outcomes are relied upon when selecting a choice.

22. A method according to claim 21 further comprising:

defining a set of the items by selecting items from the plurality of items based at least in part on the received business rules, wherein each of the plurality of adaptive agents considers the set of the items in generating the one or more recommendations.

23. A method according to claim 22, wherein the business rules include item eligibility rules which are used to select the set of items, the eligibility rules including one or more types selected from the group comprising:

geographical restrictions, customer-income restrictions, customer purchases, history restrictions, demographic restrictions, time of the day restrictions, and business condition-related restrictions.

24. A method according to claim 21, wherein the activating of one or more of the plurality of adaptive agents is carried out according to the business rules such that the at least one final recommendation originates from one of the selected adaptive agents.

25. A method according to claim 24 further comprising:

receiving a request for a recommendation; and

routing the request to a single activated adaptive agents based on the business rules.

26. A method according to claim 25, wherein the request is received from a user via a user interface.

27. A method according to claim 24, wherein the request is generated automatically.

28. A method according to claim 21, wherein the business rules include one or more business goals, and one final recommendation is selected based on the one or more business goals.

29. A method according to claim 28, wherein the one or more business goals is selected from the group comprising:

revenue maximization, profit maximization, customer satisfaction, customer retention, loss minimization, market penetration, and increase in market share.

30. A method according to claim 21 further comprising:

initiating new adaptive agents and retiring existing adaptive agents based on the received business rules.

31. A method according to claim 30, wherein the business rules include rules having one or more parameters selected from the group comprising:

how often agents are created, how many agents should be instantiated concurrently, and how long an agent can remain active before retirement.

32. A method according to claim 21 further comprising:

receiving and storing information on outcomes relating to items.

33. A method according to claim 32, wherein the information on outcomes relating to items is received from a user via a user interface.

34. A method according to claim 32 further comprising:

determining whether the information has a correlation with one or more prior recommendations.

35. A method according to claim 21, wherein the business rules are received from one or more maintenance users via a maintenance interface.

36. A method according to claim 35 further comprising:

receiving new items for inclusion in the plurality of items from one or more maintenance users via the maintenance interface.

37. A method according to claim 36 further comprising:

receiving an initial bias with respect to a new item.

38. A method according to claim 21, wherein the items include items having one or types selected from the group comprising:

products, services, offers for products, offers for services, suppliers, and agencies.

39. An article comprising a tangible machine-readable storage medium embodying instructions that when performed by one or more machines result in operations for making a recommendation with respect to a plurality of items using a plurality of adaptive agents comprising:

receiving one or more business rules;

receiving a request for a recommendation;

receiving attributes relating to the request;

activating one or more adaptive agents such that each activated adaptive agent generates one or more recommendations with respect to the one or more items based at least in part on an evaluation of prior outcomes relating to the items;

selecting from among the one or more recommendations at least one final recommendation; and

displaying the at least one final recommendation to a user;

wherein at least some of adaptive agents have variable learning rates, the learning rate influencing a degree on which prior outcomes are relied upon when selecting a choice.

40. An article according to claim 39, wherein the tangible machine-readable storage medium further embodies instructions that when performed by one or more machines result in operations comprising:

defining a set of the items by selecting items from the plurality of items based at least in part on the received business rules, wherein each of the plurality of adaptive agents considers the set of the items in generating the one or more recommendations.

41. An article according to claim 40, wherein the business rules include item eligibility rules which are used to select the set of items, the eligibility rules including one or more types selected from the group comprising:

geographical restrictions, customer-income restrictions, customer purchases, history restrictions, demographic restrictions, time of the day restrictions, and business condition-related restrictions.

42. An article according to claim 39, wherein the activating of one or more of the plurality of adaptive agents is carried out according to the business rules such that the at least one final recommendation originates from one of the selected adaptive agents.

43. An article according to claim 42, wherein the tangible machine-readable storage medium further embodies instructions that when performed by one or more machines result in operations comprising:

receiving a request for a recommendation; and

routing the request to a single activated adaptive agents based on the business rules.

44. An article according to claim 42, wherein the request is received from a user via a user interface.

45. An article according to claim 42, wherein the request is generated automatically.

46. An article according to claim 39, wherein the business rules include one or more business goals, and one final recommendation is selected based on the one or more business goals.

47. An article according to claim 46, wherein the one or more business goals is selected from the group comprising:

revenue maximization, profit maximization, customer satisfaction, customer retention, loss minimization, market penetration, and increase in market share.

48. An article according to claim 39, wherein the tangible machine-readable storage medium further embodies instructions that when performed by one or more machines result in operations comprising:

initiating new adaptive agents and retiring existing adaptive agents based on the received business rules.

49. An article according to claim 48 wherein the business rules include rules having one or more parameters selected from the group comprising:

how often agents are created, how many agents should be instantiated concurrently, and how long an agent can remain active before retirement.

50. An article according to claim 39, wherein the tangible machine-readable storage medium further embodies instructions that when performed by one or more machines result in operations comprising:

receiving and storing information on outcomes relating to items.

51. An article according to claim 50, wherein the information on outcomes relating to items is received from a user via a user interface.

52. An article according to claim 50, wherein the tangible machine-readable storage medium further embodies instructions that when performed by one or more machines result in operations comprising:

determining whether the information has a correlation with one or more prior recommendations.

53. An article according to claim 39, wherein the business rules are received from one or more maintenance users via a maintenance interface.

54. An article according to claim 53, wherein the tangible machine-readable storage medium further embodies instructions that when performed by one or more machines result in operations comprising:

receiving new items for inclusion in the plurality of items from one or more maintenance users via the maintenance interface.

55. An article according to claim 54, wherein the tangible machine-readable storage medium further embodies instructions that when performed by one or more machines result in operations comprising:

receiving an initial bias with respect to a new item.

56. An article according to claim 39, wherein the items include items having one or types selected from the group comprising:

products, services, offers for products, offers for services, suppliers, and agencies.