Generating and providing content recommendations to a group of users

• US 9,792,640 B2
• Filed: 04/28/2013
• Issued: 10/17/2017
• Est. Priority Date: 08/18/2010
• Status: Expired due to Fees

• Current Assignee: Jinni Media Limited
• Original Assignee: Jinni Media Limited
• Inventors: Meshulam, Ram, Rimon, Mordechai Mori, Ben Zaken, Izhak
• Primary Examiner(s): LEVINE, ADAM L
• Application Number:
  US13/872,117
• Publication Number:
  US 20130238710A1
• Time in Patent Office: 1,633 Days
• CPC Class Codes: G06Q 30/0605 Supply or demand aggregation
• CPC Class Codes: G06Q 30/0631 Item recommendations
• CPC Class Codes: H04L 67/306 User profiles
• CPC Class Codes: H04L 67/535 Tracking the activity of the user network monitoring arrangements H04L43/00; recording of computer activity G06F11/34

Abstract

Disclosed are systems, apparatuses, circuits, methods and computer executable code sets for generating and providing content recommendations to match the tastes and preferences of a group of users. a Recommendation Engine is used for generating two or more individual content recommendation sets for each of the members in the user group. A Recommendation Aggregation Module is used for adding and combining the individual content recommendation sets into an aggregated recommendation set. a Recommendation Selection Module is used for selecting at least a subset of the content items in the aggregated recommendation set for inclusion in a content recommendation result set. A Profile Engine is used for building individual group users profiles from which a merged group profile is constructed, or for building a single joint group profile based on inputs from multiple group users.

Claims

1. A system for generating content recommendation sets to match structured preference profiles of a group of users, said system comprising:

a first computing device including a Recommendation Engine configured to automatically generate for each given user in the group, based on a structured user preference profile of the given user;

(i) individual content recommendation sets, and (ii) a score for each content recommendation in the individual recommendation sets, the score representing a confidence of the system in each recommendation in relation to the given user;

a second computing device including a Recommendation Aggregation Module configured to combine the individual content recommendation sets into an aggregated recommendation set;

a third computing device including a relevancy level calculator configured to calculate a relevancy level of items in the aggregated recommendation set in relation to each user in the group, by calculating within a vector space defined by a defined taxonomy of media specific content genes, for each given item in the aggregated recommendation set, in relation to each given user in the group, a distance between;

(a) a vector representing a set of salience values of first semantic characteristics within the given content item, which first semantic characteristics are selected from the defined taxonomy of media specific content genes, and (b) as centroid of a vector cluster, each vector in the cluster representing a set of values of second semantic characteristics of rated content items related to the given user, which second semantic characteristics are also selected from the defined taxonomy of media specific content genes; and

a fourth computing device including a Recommendation Selection Module configured to automatically select a subset of selected content recommendations, from the aggregated recommendation set, for inclusion in a content recommendation result set for the group, based at least partially on;

a. a number of the individual content recommendation sets including the selected content recommendations;

b. the scores of the selected content recommendations in the individual content recommendation sets; and

c. the calculated relevancy levels of the selected content recommendations.

2. The system according to claim 1, further comprising a Per Item User Counter for recording a number of users to which a given content item was recommended.

3. The system according to claim 1, wherein said Recommendation Selection Module is further configured to factor a calculated similarity level above a defined threshold between one or more of the selected content recommendations and a content item associated with a recommendation history of one or more of the users, when selecting a subset of selected content recommendations.

4. The system according to claim 1, further comprising a Direct User Input Analyzer comprising processing circuitry configured to process user inputs, of users in the group, relating to content items and modify the user preference profiles accordingly.

5. The system according to claim 1, wherein said Recommendation Selection Module is further configured to factor a community rating of one or more of the selected content recommendations when selecting a subset of selected content recommendations.

6. The system according to claim 1, wherein said system is further adapted to calculate and provide per-user match rates between the selected content recommendations and each given user in the group.

7. The system according to claim 1, wherein said Recommendation Engine is functionally associated with a Content Prioritization Module for offering content items from specific preferred catalogues.

8. A method for generating content recommendation sets to match structured preference profiles of a group of users, said method comprising:

using a computing device to automatically generate for each given user in the group, based on a structured user preference profile of the given user;

(i) individual content recommendation sets, and (ii) a score for each content recommendation in the individual recommendation sets, the score representing a confidence of the system in each recommendation in relation to the given user;

using a computing device to combine the individual content recommendation sets into an aggregated recommendation set;

using a computing device to calculate a relevancy level of items in the aggregated recommendation set in relation to each given user in the group by calculating, within a vector space defined by a defined taxonomy of media specific content genes, a distance between;

(a) a vector representing a set of salience values of first semantic characteristics within the given content item, which first semantic characteristics are selected from the defined taxonomy of media specific content genes, and (b) a centroid of a vector cluster, each vector in the cluster representing a set of values of second semantic characteristics of rated content items related to the given user, which second semantic characteristics are also selected from the defined taxonomy of media specific content genes; and

using a computing device to automatically select a subset of selected content recommendations, from the aggregated recommendation set, for inclusion in a content recommendation result set for the group, based at least partially on;

a. a number of the individual content recommendation sets including the selected content recommendations;

b. the scores of the selected content recommendations in the individual content recommendation sets; and

c. the calculated relevancy levels of the selected content recommendations.

9. The method according to claim 8, wherein using a computing device to automatically select a subset of the content items further comprises recording a number of users to which a given content item was recommended.

10. The method according to claim 8, wherein using a computing device to automatically select a subset of the content items further comprises calculating and factoring a similarity level above a defined threshold between one or more of the selected content recommendations and a content item associated with a recommendation history of one or more of the users.

11. The method according to claim 8, wherein using a computing device to automatically select a subset of the content items further comprises processing user inputs, of users in the group, relating to content items and modifying the user preference profiles accordingly.

12. The method according to claim 8 wherein using a computing device to automatically select a subset of the content items further comprises factoring a Community Rating of one or more of the selected content recommendations.

13. The method according to claim 8, further comprising providing per-user match rates between the selected content recommendations and each given user in the group.

14. The method according to claim 8, wherein using a computing device to automatically select a subset of the content items includes preferring content items from specific preferred catalogues.

15. A system for generating content recommendation sets to match structured preference profiles of a group of users;

said system comprising;

a first computing device including a Recommendation Engine configured to automatically generate for each given user in the group, based on a structured user preference profile of the given user;

(i) individual content recommendation sets for the given user, and (ii) a score for each content recommendation in the individual recommendation sets, the score representing a confidence of the system in each recommendation;

a second computing device including a Recommendation Aggregation Module configured to combine the individual content recommendation sets into an aggregated recommendation set; and

a third computing device including a Recommendation Selection Module configured to automatically select a subset of selected content recommendations;

from the aggregated recommendation set, for inclusion in a content recommendation result set for the group, based at least partially on;

a. a number of the individual content recommendation sets including the selected content recommendations;

b. the scores of the selected content recommendations in the individual content recommendation sets; and

c. a community rating of one or more of the selected content recommendations

a fourth computing device including a Relevancy Level Calculator configured to calculate a relevancy level of a given item in the aggregated recommendation set, in relation to a given user in the group, by calculating, within a vector space defined by a defined taxonomy of media specific content genes, a distance between;

(a) a vector representing a set of salience values of first semantic characteristics within the given content item, which first semantic characteristics are selected from the defined taxonomy of media specific content genes, and (b) a centroid of a vector cluster, each vector in the cluster representing a set of values of second semantic characteristics of rated content items related to the given user, which second semantic characteristics are also selected from the defined taxonomy of media specific content genes; and

wherein said Recommendation Selection Module is configured to factor the calculated relevancy levels when automatically selecting a subset of selected content recommendations.

16. The system according to claim 15, wherein said Recommendation Selection Module is further configured to factor a calculated similarity level above a defined threshold between one or more of the selected content recommendations and a content item associated with a recommendation history of one or more of the users, when selecting a subset of selected content recommendations.

17. The system according to claim 15, wherein said system is further adapted to calculate and provide per-user match rates between the selected content recommendations and each given user in the group.

18. A system for generating content recommendation sets to match structured preference profiles of a group of users;

said system comprising;

a first computing device including a Recommendation Engine configured to automatically generate for each given user in the group, based on a structured user preference profile of the given user;

(i) individual content recommendation sets for the given user, and (ii) a score for each content recommendation in the individual recommendation sets, the score representing a confidence of the system in each recommendation;

a second computing device including a Recommendation Aggregation Module configured to combine the individual content recommendation sets into an aggregated recommendation set; and

a third computing device including a Recommendation Selection Module configured to automatically select a subset of selected content recommendations, from the aggregated recommendation set, for inclusion in a content recommendation result set for the group, based at least partially on;

a. a number of the individual content recommendation sets including the selected content recommendations;

b. the scores of the selected content recommendations in the individual content recommendation sets; and

c. a calculated similarity level above a defined threshold between one or more of the selected content recommendations and a content item associated with a recommendation history of one or more of the users

a fourth computing device including a Relevancy Level Calculator configured to calculate a relevancy level of a given item in the aggregated recommendation set, in relation to a given user in the group, by calculating, within a vector space defined by a defined taxonomy of media specific content genes, a distance between;

(a) a vector representing a set of salience values of first semantic characteristics within the given content item, which first semantic characteristics are selected from the defined taxonomy of media specific content genes, and (b) as centroid of a vector cluster, each vector in the cluster representing a set of values of second semantic characteristics of rated content items related to the given user, which second semantic characteristics are also selected from the defined taxonomy of media specific content genes; and

wherein said Recommendation Selection Module is configured to factor the calculated relevancy levels when automatically selecting a subset of selected content recommendations.