Apparatus, method and article to effect user interest-based matching in a network environment

• US 9,836,533 B1
• Filed: 03/25/2015
• Issued: 12/05/2017
• Est. Priority Date: 04/07/2014
• Status: Active Grant

• Current Assignee: Plentyoffish Media ULC (Match Group, Inc.)
• Original Assignee: Plentyoffish Media ULC (Match Group, Inc.)
• Inventors: Levi, Thomas S., Oldridge, Steve
• Primary Examiner(s): Stevens, Robert
• Application Number:
  US14/668,808
• Time in Patent Office: 986 Days
• Field of Search: 707739
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising

Abstract

Websites operated by relationship building service providers collect information from users for use in identifying candidate matches for a user. At least a portion of the information may include freeform text entered by the user to identify the user'"'"'s interests. The freeform text entered by users may be preprocessed and a user interest group model formed using a generative Natural Language Process that is applied to determine user interest groups from the freeform text. Such user interest groups identify related words based on the frequency of their occurrence and relationship within the freeform text. Individual user interest group membership vectors for each user included in a number of users may be determined using the user interest group model. At least one compatibility aspect between users may then be assessed based on the similarity or distance between the user interest group membership vectors associated with each user.

Claims

1. A method of operation in an evaluation system, the evaluation system including at least one processor and at least one nontransitory processor-readable medium that stores at least one of processor-executable instructions or data, the at least one nontransitory processor-readable medium communicatively coupled to the at least one processor, the method comprising:

deriving by the at least one processor a set of user interest groups from a plurality of user interests specified by a plurality of users as freeform text, where the number of interest groups in the set of user interest groups is less than a total number of user interests in the plurality of user interests specified by the plurality of users;

for each of a number of the users, determining by the at least one processor a respective interest group membership vector based on respective user interests specified by the respective user, the interest group membership vector having one or more elements which represent a partial membership in respective ones of one or more of the interest groups; and

assessing by the at least one processor at least one of the users based at least in part on at least one of the interest group membership vectors.

2. The method of claim 1 wherein assessing at least one of the users based at least in part on at least one of the interest group membership vectors includes comparing a respective interest group membership vector for a first one of the users with the respective interest group membership vector for at least one other one of the users.

3. The method of claim 2 wherein comparing a respective interest group membership vector for a first one of the users with the respective interest group membership vector for at least one other one of the users includes determining a distance between the respective interest group membership vector for the first one of the users and the respective interest group membership vector for the at least one other one of the users.

4. The method of claim 2 wherein comparing a respective interest group membership vector for a first one of the users with the respective interest group membership vector for at least one other one of the users includes determining a similarity between the respective interest group membership vector for the first one of the users and the respective interest group membership vector for the at least one other one of the users.

5. The method of claim 1, further comprising:

for at least one of the interest group membership vectors, normalizing by the at least one processor the respective interest group membership vector, and wherein assessing at least one of the users based at least in part on at least one of the interest group membership vectors includes assessing at least a first one of the users based at least in part on the normalized respective interest group membership vector for the first one of the users.

6. The method of claim 1 wherein for each of a number of the users, determining a respective interest group membership vector based on respective user interests specified by the respective user includes determining a respective interest group membership vector for each active user associated with a first defined language or first defined dialect of a language.

7. The method of claim 1, further comprising:

receiving the plurality of user interests specified by the plurality of users as freeform text; and

preprocessing the received plurality of user interests specified by the plurality of users as freeform text.

8. The method of claim 7 wherein preprocessing the received plurality of user interests specified by the plurality of users as freeform text includes at least one of:

removing by the at least one processor characters that are not valid letters in a valid language from the received plurality of user interests specified by the plurality of users as freeform text;

removing by the at least one processor stop words from the received plurality of user interests specified by the plurality of users as freeform text;

or

performing by the at least one processor stemming on the received plurality of user interests specified by the plurality of users as freeform text.

9. The method of claim 8 wherein deriving a set of user interest groups from a plurality of user interests specified by a plurality of users as freeform text includes deriving the set of interest groups from the preprocessed plurality of user interests specified by the plurality of users, where user interests includes at least one incorrect spelling of a user interest, and at least one pair of alternative names for a given user interest.

10. The method of claim 1 wherein deriving a set of user interest groups from a plurality of user interests specified by a plurality of users as freeform text includes generating a generative natural language processing model based at least in part on the plurality of user interests specified by the plurality of users as freeform text.

11. The method of claim 10 wherein generating a generative natural language processing model includes preforming a Latent Dirichlet Allocation on preprocessed user interests specified by the plurality of users as freeform text.

12. The method of claim 10 wherein each of the users provides a respective set of user interest words in freeform text and deriving a set of user interest groups from a plurality of user interests specified by a plurality of users as freeform text includes computing for a number of interest words a respective term frequency-inverse document frequency.

13. The method of claim 1, further comprising:

receiving a set of self-specified user interests provided by a first one of the users, and wherein assessing at least one of the users based at least in part on at least one of the interest group membership vectors includes assessing potential matches with other ones of the users based on an assessment in commonality or differences between the respective interest group membership vector of the first one of the users and respective interest group membership vectors of the other ones of the users.

14. The method of claim 1, further comprising:

receiving a set of user interests desired by a first one of the users in another one of the users;

generating a hypothetical interest group membership vector based on the received set of user interests desired by the first one of the users in another one of the users, and wherein assessing at least one of the users based at least in part on at least one of the interest group membership vectors includes assessing potential matches with other ones of the users based on an assessment in commonality or differences between the hypothetical interest group membership vector and respective interest group membership vectors of the other ones of the users.

15. The method of claim 1, further comprising:

receiving an identification of one user by a first one of the users, and wherein assessing at least one of the users based at least in part on at least one of the interest group membership vectors includes assessing potential matches with other ones of the users based on an assessment in commonality or differences between a respective interest group membership vector of the identified one of the users and respective interest group membership vectors of other ones of the users.

16. The method of claim 15 wherein receiving an identification of one user by a first one of the users includes receiving identification of at least one user who has at least one trait, other than user interests, which disqualifies the respective at least one user as a match for the first one of the users.

17. The method of claim 1, further comprising:

detecting an interest in one user by a first one of the users, and wherein assessing at least one of the users based at least in part on at least one of the interest group membership vectors includes assessing potential matches with other ones of the users based on an assessment in commonality or differences between a respective interest group membership vector of the one of the users in which the interest was detected and respective interest group membership vectors of other ones of the users.

18. The method of claim 1, further comprising:

selecting by the at least one processor a first subset of the users based on traits other than user interests, and wherein assessing at least one of the users based at least in part on at least one of the interest group membership vectors includes assessing a compatibility of the users of the first subset of users with a first one of the users based on a respective interest group membership vector for the first one of the users with the respective interest group membership vector for the other one of the users in the first subset of the users.

19. The method of claim 1, further comprising:

determining by the at least one processor at least one hypothetical interest group membership vector based on the respective interest group membership vectors of a plurality of the users, and wherein assessing at least one of the users based at least in part on at least one of the interest group membership vectors includes comparing a respective interest group membership vector for a first user with the at least one hypothetical interest group membership vector.

20. The method of claim 19 wherein determining at least one hypothetical interest group membership vector based on the respective interest group membership vectors of a plurality of the users includes determining at least one of a mean or median of the respective interest group membership vectors of the set of users.

21. The method of claim 19 wherein determining at least one hypothetical interest group membership vector based on the respective interest group membership vectors of a plurality of the users includes determining the at least one hypothetical interest group membership vector based on the respective interest group membership vectors of a set of users that each have at least one demographic attribute in common.

22. The method of claim 19 wherein determining at least one hypothetical interest group membership vector based on the respective interest group membership vectors of a plurality of the users includes determining the at least one hypothetical interest group membership vector based on the respective interest group membership vectors of a set of users that each have a geographic locale in common.

23. The method of claim 19, further comprising:

determining a relative distance between the respective interest group membership vector of the first user and the at least one hypothetical interest group membership vector; and

providing an indication of the determined relative distance.

24. The method of claim 19, further comprising:

identifying at least one other set of users with which the first user is more compatible based on the respective interest group membership vector of the first user and at least one hypothetical interest group membership vector of the at least one other set of users; and

providing an indication of the determined other set of users with which the first user is more compatible.

25. The method of claim 24, further comprising:

selecting the set of users based on at least a first characteristic that is common to the users of the first set; and

selecting the at least one other set of users based on at least a second characteristic that is common to the users of the second set and not common to the users of the first set.

26. The method of claim 24, further comprising:

selecting the set of users based on at least a first geographical locale that is common to the users of the first set; and

selecting the at least one other set of users based on at least a second geographic locale that is common to the users of the second set and not common to the users of the first set.

27. The method of claim 1, further comprising:

selecting by the at least one processor a first subset of the users based on traits other than user interests, and wherein assessing at least one of the users based at least in part on at least one of the interest group membership vectors includes assessing a dissimilarity of the users of the first subset of users with a first one of the users based on a respective interest group membership vector for the first one of the users with the respective interest group membership vector for the other one of the users in the first subset of the users.

28. The method of claim 1, further comprising:

determining by the at least one processor at least one hypothetical interest group membership vector based on the respective interest group membership vectors of a plurality of the users, and wherein assessing at least one of the users based at least in part on at least one of the interest group membership vectors includes displaying at least one interest-based characteristic of the group identified at least in part via the hypothetical interest group membership vector.

29. The method of claim 1, further comprising:

determining by the at least one processor a first hypothetical interest group membership vector based on the respective interest group membership vectors of a first plurality of the users;

determining by the at least one processor a second hypothetical interest group membership vector based on the respective interest group membership vectors of a second plurality of the users; and

wherein assessing at least one of the users based at least in part on at least one of the interest group membership vectors includes determining at least one of a similarity or a dissimilarity between the first plurality of the users the second plurality of the users by comparing the first hypothetical interest group membership vector with the second hypothetical group membership vector.

30. The method of claim 1, further comprising:

determining by the at least one processor the first hypothetical interest group membership vector based on the respective interest group membership vectors of the first plurality of the users who share at least a first non-interest based trait;

determining by the at least one processor a second hypothetical interest group membership vector based on the respective interest group membership vectors of the second plurality of the users who share at least a second non-interest based trait that is different than the first non-interest based trait; and

wherein assessing at least one of the users based at least in part on at least one of the interest group membership vectors includes comparing the interests of the first plurality of the users as provided by the first hypothetical interest group membership vector with the interests of the second plurality of the users as provided by the second hypothetical group membership vector.

31. The method of claim 30 wherein determining by the at least one processor the first hypothetical interest group membership vector based on the respective interest group membership vectors of the first plurality of the users who share at least a first non-interest based trait includes;

determining by the at least one processor the first hypothetical interest group membership vector based on the respective interest group membership vectors of the first plurality of the users who share at least a first non-interest based trait that includes a defined first geographic location; and

wherein determining by the at least one processor the second hypothetical interest group membership vector based on the respective interest group membership vectors of the second plurality of the users who share at least the second non-interest based trait includes;

determining by the at least one processor the second hypothetical interest group membership vector based on the respective interest group membership vectors of the second plurality of the users who share at least the second non-interest based trait that includes a defined second geographic location.

32. The method of claim 30 wherein the first plurality of the users and the second plurality of the users share at least a third non-interest based trait and wherein determining by the at least one processor the first hypothetical interest group membership vector based on the respective interest group membership vectors of the first plurality of the users who share at least a first non-interest based trait and a second non-interest based trait includes;

determining by the at least one processor the first hypothetical interest group membership vector based on the respective interest group membership vectors of the first plurality of the users who share at least a first non-interest based trait that includes a defined first geographic location and a third non-interest based trait; and

wherein determining by the at least one processor the second hypothetical interest group membership vector based on the respective interest group membership vectors of the second plurality of the users who share at least the second non-interest based trait and a third non-interest based trait includes;

determining by the at least one processor the second hypothetical interest group membership vector based on the respective interest group membership vectors of the second plurality of the users who share at least the second non-interest based trait that includes a defined second geographic location and a third non-interest based trait.

33. An evaluation system, comprising:

at least one processor; and

at least one nontransitory processor-readable medium communicatively coupled to the at least one processor that stores at least one of processor-executable instructions or data that when executed by the at least one processor cause the at least one processor to;

derive a set of user interest groups from a plurality of user interests specified by a plurality of users as freeform text, where the number of interest groups in the set of user interest groups is less than a total number of user interests in the plurality of user interests specified by the plurality of users;

for each of a number of the users, determine a respective interest group membership vector based on respective user interests specified by the respective user the interest group membership vector having one or more elements which represent a partial membership in respective ones of one or more of the interest groups; and

assess at least one of the users based at least in part on at least one of the interest group membership vectors.

34. The system of claim 33 wherein the processor-executable instructions that cause the at least one processor to assess at least one of the users based at least in part on at least one of the interest group membership vectors, further cause the at least one processor to:

compare a respective interest group membership vector for a first one of the users with the respective interest group membership vector for at least one other one of the users.

35. The system of claim 34 wherein the processor-executable instructions that cause the at least one processor to compare the respective interest group membership vector for a first one of the users with the respective interest group membership vector for at least one other one of the users, further causes the at least one processor to:

determine a distance between the respective interest group membership vector for the first one of the users and the respective interest group membership vector for the at least one other one of the users.

36. The system of claim 34 wherein the processor-executable instructions that cause the at least one processor to compare a respective interest group membership vector for a first one of the users with the respective interest group membership vector for at least one other one of the users, further causes the at least one processor to:

determine a similarity between the respective interest group membership vector for the first one of the users and the respective interest group membership vector for the at least one other one of the users.

37. The system of claim 33 wherein the processor-executable instructions further cause the at least one processor to:

normalize the respective interest group membership vector for at least one of the interest group membership vectors; and

wherein the processor-executable instructions that cause the at least one processor to assess at least one of the users based at least in part on at least one of the interest group membership vectors, further causes the at least one processor to;

assess at least a first one of the users based at least in part on the normalized respective interest group membership vector for the first one of the users.

38. The system of claim 33 wherein the processor-executable instructions that cause the at least one processor to determine a respective interest group membership vector based on respective user interests specified by the respective user, further causes the at least one processor to:

determine a respective interest group membership vector for each active user associated with a first defined language or first defined dialect of a language.

39. The system of claim 33 wherein the processor-executable instructions further cause the at least one processor to:

receive the plurality of user interests specified by the plurality of users as freeform text; and

preprocess the received plurality of user interests specified by the plurality of users as freeform text.

40. The system of claim 39 wherein the processor-executable instructions that cause the at least one processor to preprocess the received plurality of user interests specified by the plurality of users as freeform text, further causes the at least one processor to at least one of:

remove characters that are not valid letters in a valid language from the received plurality of user interests specified by the plurality of users as freeform text;

remove stop words from the received plurality of user interests specified by the plurality of users as freeform text;

or

perform stemming on the received plurality of user interests specified by the plurality of users as freeform text.

41. The system of claim 40 wherein the processor-executable instructions that cause the at least one processor to derive a set of user interest groups from a plurality of user interests specified by a plurality of users as freeform text, further causes the at least one processor to:

derive the set of interest groups from the preprocessed plurality of user interests specified by the plurality of users, which user interests includes at least one incorrect spelling of a user interest, and at least one pair of alternative names for a given user interest.

42. The system of claim 33 wherein the processor-executable instructions that cause the at least one processor to derive a set of user interest groups from a plurality of user interests specified by a plurality of users as freeform text, further cause the at least one processor to:

generate a generative natural language processing model based at least in part on the plurality of user interests specified by the plurality of users as freeform text.

43. The system of claim 42 wherein the processor-executable instructions that cause the at least one processor to generate a generative natural language processing model, further cause the at least one processor to:

perform a Latent Dirichlet Allocation on preprocessed user interests specified by the plurality of users as freeform text.

44. The system of claim 42 wherein the processor-executable instructions that cause the at least one processor to derive a set of user interest groups from a plurality of user interests specified by a plurality of users as freeform text, further causes the at least one processor to:

compute for a number of interest words, provided by each of the users provides a respective set of user interest words in freeform text, a respective term frequency-inverse document frequency.

45. The system of claim 33 wherein the processor-executable instructions further cause the at least one processor to:

receive a set of self-specified user interests provided by a first one of the users, and wherein assessing at least one of the users based at least in part on at least one of the interest group membership vectors includes assessing potential matches with other ones of the users based on an assessment in commonality or differences between the respective interest group membership vector of the first one of the users and respective interest group membership vectors of the other ones of the users.

46. The system of claim 33 wherein the processor-executable instructions further cause the at least one processor to:

receive a set of user interests desired by a first one of the users in another one of the users;

generate a hypothetical interest group membership vector based on the received set of user interests desired by the first one of the users in another one of the users, and wherein assessing at least one of the users based at least in part on at least one of the interest group membership vectors includes assessing potential matches with other ones of the users based on an assessment in commonality or differences between the hypothetical interest group membership vector and respective interest group membership vectors of the other ones of the users.

47. The system of claim 33 wherein the processor-executable instructions further cause the at least one processor to:

receive an identification of one user by a first one of the users, and wherein assessing at least one of the users based at least in part on at least one of the interest group membership vectors includes assessing potential matches with other ones of the users based on an assessment in commonality or differences between a respective interest group membership vector of the identified one of the users and respective interest group membership vectors of other ones of the users.

48. The system of claim 47 wherein the processor-executable instructions that cause the at least one processor to receive an identification of one user by a first one of the users, further causes the at least one processor to:

receive identification of at least one user who has at least one trait, other than user interests, which disqualifies the respective at least one user as a match for the first one of the users.

49. The system of claim 33 wherein the processor-executable instructions further cause the at least one processor to:

detect an interest in one user by a first one of the users, and wherein assessing at least one of the users based at least in part on at least one of the interest group membership vectors includes assessing potential matches with other ones of the users based on an assessment in commonality or differences between a respective interest group membership vector of the one of the users in which the interest was detected and respective interest group membership vectors of other ones of the users.

50. The system of claim 33 wherein the processor-executable instructions further cause the at least one processor to:

select by the at least one processor a first subset of the users based on traits other than user interests, and wherein assessing at least one of the users based at least in part on at least one of the interest group membership vectors includes assessing a compatibility of the users of the first subset of users with a first one of the users based on a respective interest group membership vector for the first one of the users with the respective interest group membership vector for the other one of the users in the first subset of the users.

51. The system of claim 33 wherein the processor-executable instructions further cause the at least one processor to:

determine by the at least one processor at least one hypothetical interest group membership vector based on the respective interest group membership vectors of a plurality of the users, and wherein assessing at least one of the users based at least in part on at least one of the interest group membership vectors includes comparing a respective interest group membership vector for a first user with the at least one hypothetical interest group membership vector.

52. The system of claim 51 wherein the processor-executable instructions that cause the at least one processor to determine at least one hypothetical interest group membership vector based on the respective interest group membership vectors of a plurality of the users, further causes the at least one processor to:

determine at least one of a mean or median of the respective interest group membership vectors of the set of users.

53. The system of claim 51 wherein the processor-executable instructions that cause the at least one processor to determine at least one hypothetical interest group membership vector based on the respective interest group membership vectors of a plurality of the users, further causes the at least one processor to:

determine the at least one hypothetical interest group membership vector based on the respective interest group membership vectors of a set of users that each have at least two demographic attributes in common.

54. The system of claim 51 wherein the processor-executable instructions that cause the at least one processor to determine at least one hypothetical interest group membership vector based on the respective interest group membership vectors of a plurality of the users, further causes the at least one processor to:

determine the at least one hypothetical interest group membership vector based on the respective interest group membership vectors of a set of users that each have a geographic locale in common.

55. The system of claim 51 wherein the processor-executable instructions further cause the at least one processor to:

determine a relative distance between the respective interest group membership vector of the first user and the at least one hypothetical interest group membership vector; and

provide an indication of the determined relative distance.

56. The system of claim 51 wherein the processor-executable instructions further cause the at least one processor to:

identify at least one other set of users with which the first user is more compatible based on the respective interest group membership vector of the first user and at least one hypothetical interest group membership vector of the at least one other set of users; and

provide an indication of the determined other set of users with which the first user is more compatible.

57. The system of claim 56 wherein the processor-executable instructions further cause the at least one processor to:

select the set of users based on at least a first characteristic that is common to the users of the first set; and

select the at least one other set of users based on at least a second characteristic that is common to the users of the second set and not common to the users of the first set.

58. The system of claim 56 wherein the processor-executable instructions further cause the at least one processor to:

select the set of users based on at least a first geographical locale that is common to the users of the first set; and

select the at least one other set of users based on at least a second geographic locale that is common to the users of the second set and not common to the users of the first set.

59. The system of claim 33 wherein the processor-executable instructions further cause the at least one processor to:

select a first subset of the users based on traits other than user interests; and

wherein the processor-executable instructions that cause the at least one processor to assess at least one of the users based at least in part on at least one of the interest group membership vectors, further cause the at least one processor to;

assess a dissimilarity of the users of the first subset of users with a first one of the users based on a respective interest group membership vector for the first one of the users with the respective interest group membership vector for the other one of the users in the first subset of the users.

60. The system of claim 33 wherein the processor-executable instructions further cause the at least one processor to:

determine at least one hypothetical interest group membership vector based on the respective interest group membership vectors of a plurality of the users;

wherein the processor-executable instructions that cause the at least one processor to assess at least one of the users based at least in part on at least one of the interest group membership vectors, further cause the at least one processor to;

display at least one interest-based characteristic of the group identified at least in part via the hypothetical interest group membership vector.

61. The system of claim 33 wherein the processor-executable instructions further cause the at least one processor to:

determine a first hypothetical interest group membership vector based on the respective interest group membership vectors of a first plurality of the users; and

determine a second hypothetical interest group membership vector based on the respective interest group membership vectors of a second plurality of the users;

wherein the processor-executable instructions that cause the at least one processor to assess at least one of the users based at least in part on at least one of the interest group membership vectors, further cause the at least one processor to;

assess at least one of a similarity or a dissimilarity between the first plurality of the users the second plurality of the users by comparing the first hypothetical interest group membership vector with the second hypothetical group membership vector.

62. The system of claim 33 wherein the processor-executable instructions further cause the at least one processor to:

determine the first hypothetical interest group membership vector based on the respective interest group membership vectors of the first plurality of the users who share at least a first non-interest based trait and a second non-interest based trait; and

determine a second hypothetical interest group membership vector based on the respective interest group membership vectors of the second plurality of the users who share at least the first non-interest based trait and a third non-interest based trait that is different than the second non-interest based trait;

wherein the processor-executable instructions that cause the at least one processor to assess at least one of the users based at least in part on at least one of the interest group membership vectors, further cause the at least one processor to;

assess the interests of the first plurality of the users as provided by the first hypothetical interest group membership vector with the interests of the second plurality of the users as provided by the second hypothetical group membership vector.

63. The system of claim 62 wherein the processor-executable instructions that cause the at least one processor to determine the first hypothetical interest group membership vector based on the respective interest group membership vectors of the first plurality of the users who share at least a first non-interest based trait and a second non-interest based trait, further cause the at least one processor to:

determine the first hypothetical interest group membership vector based on the respective interest group membership vectors of the first plurality of the users who share at least a first non-interest based trait and a second non-interest based trait that includes a defined first geographic location; and

wherein the processor-executable instructions that cause the at least one processor to determine by the at least one processor the second hypothetical interest group membership vector based on the respective interest group membership vectors of the second plurality of the users who share at least the first non-interest based trait and a third non-interest based trait further cause the at least one processor to;

determine the second hypothetical interest group membership vector based on the respective interest group membership vectors of the second plurality of the users who share at least the first non-interest based trait and a third non-interest based trait that includes a defined second geographic location.

64. The system of claim 62 wherein the processor-executable instructions that cause the at least one processor to determine by the at least one processor the first hypothetical interest group membership vector based on the respective interest group membership vectors of the first plurality of the users who share at least a first non-interest based trait and a second non-interest based trait further cause the at least one processor to:

determine by the at least one processor the first hypothetical interest group membership vector based on the respective interest group membership vectors of the first plurality of the users who share at least a first non-interest based trait that includes a defined first geographic location and a third non-interest based trait; and

wherein the processor-executable instructions that cause the at least one processor to determine the second hypothetical interest group membership vector based on the respective interest group membership vectors of the second plurality of the users who share at least the second non-interest based trait and a third non-interest based trait, further cause the at least one processor to;

determine the second hypothetical interest group membership vector based on the respective interest group membership vectors of the second plurality of the users who share at least the second non-interest based trait that includes a defined second geographic location and the third non-interest based trait.