System, method, and article of manufacture for making a compatibility-aware recommendations to a user

• US 6,412,012 B1
• Filed: 12/23/1998
• Issued: 06/25/2002
• Est. Priority Date: 12/23/1998
• Status: Expired due to Term

• Current Assignee: CXT Systems, Inc. (Quest Patent Research Corporation)
• Original Assignee: Net Perceptions, Inc. (Stamford Industrial Group, Inc.)
• Inventors: Frankowski, Dan, Rauser, John, Konstan, Joseph, Bieganski, Paul
• Primary Examiner(s): WILEY, DAVID ARMAND
• Application Number:
  US09/219,585
• Time in Patent Office: 1,280 Days
• Field of Search: 709/232-234, 705/26-27, 705/10, 235/383, 395/200.49, 710/62
• US Class Current: 709/232
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising

Abstract

A recommendation process includes the consideration of the compatibility of the items being recommended. An electronic processing system for generating a compatibility-aware recommendation output set to a is user based, at least in part, on a set of item compatibility rules. The system includes a processing system of one or more processors configured to receive applicable data, including i) user preference data, and ii) item compatibility rules, and to produce a compatibility-aware recommendation output set using the user preference data and the item compatibility rules. A method and computer-readable storage device containing the method are also presented.

Claims

1. An electronic processing system for generating a compatibility-aware recommendation output set to a user based, at least in part, on a set of item compatibility rules, the system comprising a processing system of one or more processors configured to:

a. receive applicable data including

i. user preference data, and

ii. item compatibility rules, and

b. produce a compatibility-aware recommendation output set using the user preference data and the item compatibility rules.

2. A system as recited in claim 1, wherein the processing system is further configured to

a. receive match data, and

b. produce the compatibility-aware recommendation output set, using the user preference data, the match data and the compatibility rules.

3. A system as recited in claim 1, further comprising a ratings interface operatively coupled to the processing system, to receive the user preference data and transmit the user preference data to the processing system.

4. A system as recited in claim 3, wherein

the ratings interface is further configured to receive user item preference data that includes at least one of unary values, binary values, and numerical values, and

the processing system is further configured to produce the compatibility-aware item recommendation set from the at least one of the unary values, binary values and numerical values.

6. A system as recited in claim 1, further comprising

an output interface operatively coupled to the processing system in order to receive the compatibility-aware recommendation output set, and

a display device operatively coupled to the output interface in order to display the compatibility-aware recommendation output set.

7. A system as recited in claim 1, wherein the processing system is further configured to derive the item compatibility rules from one or more of:

a. a marketer specification,

b. automatic generation using machine learning,

c. automatic generation using a genetic algorithm,

d. automatic generation using a neural network,

e. automatic generation using a rule inference system,

f. data mining,

g. an analysis of historical purchase and preference data, and

h. a user specification.

8. A system as recited in claim 1, wherein the processing system is further configured to receive the item compatibility rules represented as at least one of:

a. unidirectional rules,

b. bidirectional rules,

c. generalized rules including multi-way rules,

d. rules among items,

e. rules among sets,

f. rules among collections,

g. rules with weight factors,

h. rules with priorities, and

i. unweighted and unprioritized rules.

9. A system as recited in claim 1, wherein the processing system is further configured to

a. determine whether a match set exists,

b. identify items compatible with items in the match set when the match set is determined to exist, where the items compatible with items in the match set are not substitutes for items in the match set,

c. obtain recommendation scores for the identified items,

d. recommend to the user a subset of the identified compatible items that includes no items that are substitutes for other items in the subset.

10. A system as recited in claim 9, further configured to present recommended compatible items in the subset to the user with an indicator of recommendation strength for the items in the subset.

11. A system as recited in claim 1, wherein the processing system is further configured to:

a. initialize the compatibility-aware recommendation set,

b. determine whether a match set exists,

c. for each particular item in the match set, when the match set is determined to exist, apply rules having the particular item in the match set on the rules'"'"' inputs, and add items to the compatibility-aware recommendation set and subtract items from the compatibility-aware recommendation set when the applied rules so determine,

d. sort items in the compatibility-aware recommendation output set according to recommendation value,

e. for each examined item whose recommendation score exceeds a recommendation value, select a first sub-set of rules whose inputs include the examined item,

f. select a second sub-set of rules from the first sub-set of rules where rules in the second sub-set have outputs with items not appearing above the examined item in the compatibility-aware output set,

g. apply the rules in the second sub-set,

h. sort items in the compatibility-aware recommendation output set below the examined item according to recommendation score, and

i. sort all items in the compatibility-aware recommendation output set according to recommendation score.

12. A system as recited in claim 11, wherein applying a rule to a selected item includes

examining a rule having the selected item on an input to the rule,

adding an item on an output of the rule to the compatibility-aware recommendation output set where the item on output of the rule is not already in the compatibility-aware recommendation output set, and applying a neutral recommendation score to the added item, and

adding a recommendation score modifier to a recommendation score for all items on the output of the rule.

13. A system as claimed in claim 1, further configured to:

a. determine that a first and a second item have each been selected a number of times that is greater than a selection threshold,

b. determine the number of times the first item has been selected by the user when the user has selected the second item, and

c. compare, for one of the first and second items, the number of times that the one of the first and second items has been selected at the same time as the other of the first and second items, with the total number of times the one of the first and second items has been selected to produce a coincidence indicator, and

d. infer a substitute rule between the two items when the coincidence indicator indicates coincident selection of the first and second items at a level below a substitute threshold level.

14. A system as claimed in claim 1, further configured to:

a. determine that a first and a second item have each been selected a number of times that is greater than a selection threshold,

b. determine the number of times the first item has been selected by the user when the user has selected the second item, and

c. compare, for one of the first and second items, the number of times that the one of the first and second items has been selected at the same time as the other of the first and second items, with the total number of times the one of the first and second items has been selected to produce a coincidence indicator; and

d. infer a complementary rule between the two items when the coincidence indicator indicates coincident selection of the first and second items at a level above a coincidence threshold level.

15. A system as recited in claim 1, wherein the processing system is further configured to produce the compatibility-ware recommendation output set under real-time, interactive time-constraints.

16. A system as recited in claim 1, wherein the processing system is further configured to produce the compatibility-aware recommendation output set as at least one of unary, unordered and priority-ordered recommendations.

17. A system as recited in claim 1, wherein the processing system includes a first set of processors configured to

a. receive the applicable data including

i. the user preference data, and

ii. the item compatibility rules, and

b. produce the compatibility-aware recommendation output set using the user preference data and the item compatibility rules.

18. A system as recited in claim 1, wherein the processing system includes

a. a first set of processors configured to receive the user preference data,

b. a second set of processors configured to receive the item compatibility rules, and

c. a third set of processors operatively coupled to the first and second sets of processors to receive the user preference data and the item compatibility respectively therefrom and configured to produce the compatibility-aware recommendation output set using the user preference data and the item compatibility rules.

19. A system as recited in claim 1, wherein the processing system includes

a. a first set of processors configured to receive the user preference data, and

b. a second set of processors operatively coupled to the first set of processors to receive the user preference data from the first set of processors, and configured to

i) receive the item compatibility rules, and

ii) produce the compatibility-aware recommendation output set using the user preference data and the item compatibility rules.

5. A system as recited in 1, further comprising

a recommendation request interface operatively coupled to the processing system to receive a recommendation request from the user, and

wherein the processing system is configured to produce the compatibility-aware recommendation set in response to the recommendation request received from the user.

20. A method of producing a compatibility filtered and weighted recommendation to a user, the method using a computer having a processing system having one or more processors and an input/output interface, the method comprising:

a. receiving applicable data, using the processing system, including

i. user preference data, and

ii. item compatibility rules, and

b. producing, using the processing system, a compatibility-aware recommendation output set using the user preference data and the item compatibility rules.

21. A method as recited in claim 20, further comprising

a. receiving match data, using the processing system, and

b. producing, using the processing system, the compatibility-aware recommendation output set using the user preference data, the match data and the compatibility rules.

22. A method as recited in claim 20, further comprising receiving the user preference data by the interface and transmitting the user preference data to the processing system.

23. A method as recited in claim 22, further comprising

receiving, using the interface, user item preference data that includes at least one of unary values, binary values, and numerical values, and

producing the compatibility-aware item recommendation set, using the processing system, from the at least one of the unary values, binary values and numerical values.

25. A method as recited in claim 20, further comprising outputting the compatibility aware recommendation set to a display device via the input/output interface.

26. A method as recited in claim 20, further comprising deriving, using the processor system, the item compatibility rules from one or more of:

a. a marketer specification,

b. automatic generation using machine learning,

c. automatic generation using a genetic algorithm,

d. automatic generation using a neural network,

e. automatic generation using a rule inference system,

f. data mining,

g. an analysis of historical purchase and preference data, and

h. a user specification.

27. A system as recited in claim 20, further comprising receiving, using the processing system, the item compatibility rules represented as at least one of:

a. unidirectional rules,

b. bidirectional rules,

c. generalized rules including multi-way rules,

d. rules among items,

e. rules among sets,

f. rules among collections,

g. rules with weight factors,

h. rules with priorities, and

i. unweighted and unprioritized rules.

28. A method as recited in claim 20, further comprising

a. determining, using the processing system, whether a match set exists,

b. identifying items compatible with items in the match set when the match set is determined to exist, where the items compatible with items in the match set are not substitutes for items in the match set, using the processing system,

c. obtaining recommendation scores for the identified items, using the processing system, and

d. recommending to the user, using the processing system, a subset of the identified compatible items that includes no items that are substitutes for other items in the subset.

29. A method as recited in claim 28, further comprising presenting recommended compatible items in the subset to the user with an indicator of recommendation strength for the items in the subset, using the processing system.

30. A method as recited in claim 20, further comprising

a. initializing the compatibility-aware recommendation set, using the processing system,

b. determining, using the processing system, whether a match set exists,

c. for each particular item in the match set, when the match set is determined to exist, applying rules having the particular item in the match set on the rules'"'"' inputs, and adding items to the compatibility-aware recommendation set and subtracting items from the compatibility-aware recommendation set when the applied rules so determine, using the processing system,

d. sorting, using the processing system, items in the compatibility-aware recommendation output set according to recommendation value,

e. for each examined item whose recommendation score exceeds a recommendation value, selecting a first sub-set of rules whose inputs include the examined item, using the processing system,

f. selecting, using the processing system, a second sub-set of rules from the first sub-set of rules where rules in the second sub-set have outputs with items not appearing above the examined item in the compatibility-aware output set,

g. applying the rules in the second sub-set, using the processing system,

h. sorting, using the processing system, items in the compatibility-aware recommendation output set below the examined item according to recommendation score, and

i. sorting all items in the compatibility-aware recommendation output set according to recommendation score, using the processing system.

31. A method as recited in claim 30, wherein applying a rule to a selected item includes

examining a rule having the selected item on an input to the rule, using the processing system,

adding an item on an output of the rule to the compatibility-aware recommendation output set where the item on the output of the rule is not already in the compatibility-aware recommendation output set, and applying a neutral recommendation score to the added item, using the processing system, and

adding a recommendation score modifier to a recommendation score for all items on the output of the rule, using the processing system.

32. A method as claimed in claim 20, further comprising

a. determining, using the processing system, that a first and a second item have each been selected a number of times that is greater than a selection threshold,

b. determining, using the processing system, the number of times the first item has been selected by the user when the user has selected the second item,

c. comparing, for one of the first and second items, the number of times that the one of the first and second items has been selected at the same time as the other of the first and second items, with the total number of times the one of the first and second items has been selected to produce a coincidence indicator, using the processing system, and

d. inferring a substitute rule, using the processing system, between the two items when the coincidence indicator indicates coincident selection of the first and second items at a level below a substitute threshold level.

33. A method as claimed in claim 20, further comprising

a. determining, using the processing system, that a first and a second item have each been selected a number of times that is greater than a selection threshold,

b. determining, using the processing system, the number of times the first item has been selected by the user when the user has selected the second item,

c. comparing, for one of the first and second items, the number of times that the one of the first and second items has been selected at the same time as the other of the first and second items, with the total number of times the one of the first and second items has been selected to produce a coincidence indicator, using the processing system, and

d. inferring a complementary rule, using the processing system, between the two items when the coincidence indicator indicates coincident selection of the first and second items at a level above a coincidence threshold level.

34. A method as recited in claim 20, further comprising producing the compatibility-aware recommendation output set under real-time, interactive time-constraints, using the processing system.

35. A method as recited in claim 20, further comprising producing, using the processing system, the compatibility-aware recommendation output set as at least one of unary, unordered and priority-ordered recommendations.

36. A method as recited in claim 20, wherein the processing system includes a first set of processors, and further comprising

a. receiving, with the first set of processors, the applicable data including

i. the user preference data, and

ii. the item compatibility rules, and

b. producing, with the first set of processors, the compatibility-aware recommendation output set using the user preference data and the item compatibility rules.

37. A method system as recited in claim 20, wherein the processing system includes first, second and third sets of processors operatively coupled together, and further comprising

a. receiving the user preference data using the first set of processors,

b. receiving the item compatibility rules using the second set of processors, and

c. receiving, using the third set of processors, the user preference data and the item compatibility respectively from the first and second sets of processors, and producing the compatibility-aware recommendation output set using the user preference data and the item compatibility rules.

38. A method as recited in claim 20, wherein the processing system includes first and second sets of operatively coupled processors, and further comprising

a. receiving the user preference data using the first set of processors and

b. receiving, using the second set of processors, the user preference data from the first set of processors, and

i) receiving, using the second set of processors, the item compatibility rules, and

ii) producing, using the second set of processors, the compatibility-aware recommendation output set using the user preference data and the item compatibility rules.

24. A method as recited in 20, further comprising

receiving a recommendation request from the user via the interface, and

producing, using the processing system, the compatibility-aware recommendation set in response to the recommendation request received from the user.

39. A computer-readable program storage device, having a set of program instructions physically embodied thereon, executable by a computer, to perform a method of producing a compatibility-aware recommendation, the method comprising:

a. receiving applicable data including

i. user preference data, and

ii. item compatibility rules, and

b. producing a compatibility-aware recommendation output set using the user preference data and the item compatibility rules.

40. A storage device as recited in claim 39, the method further comprising

a. receiving match data; and

b. producing the compatibility-aware recommendation output set, using the user preference data, the match data and the compatibility rules.

42. A storage device as recited in claim 39, further comprising

receiving the compatibility-aware recommendation output set, and

displaying the compatibility-aware recommendation output set to the user.

41. A storage device as recited in 39, further comprising

receiving a recommendation request from the user and

producing the compatibility-aware recommendation set in response to the recommendation request received from the user.