System and method for real-time generation of segments for a computerized decision system

• US 10,733,626 B2
• Filed: 09/29/2017
• Issued: 08/04/2020
• Est. Priority Date: 09/30/2016
• Status: Active Grant

• Current Assignee: Disney Enterprises Incorporated (The Walt Disney Company)
• Original Assignee: Disney Enterprises Incorporated (The Walt Disney Company)
• Inventors: Somaya, Vikram, Srivastava, Ajit Kumar, Parr, Molly, Coleman, Sophie M., Chattleton, Christen Johnson, Keating, Scott T.
• Primary Examiner(s): Li, Sun M
• Application Number:
  US15/719,866
• Publication Number:
  US 20180096385A1
• Time in Patent Office: 1,040 Days
• Field of Search: 705 1449, 705 1473, 705 37, 705 144
• CPC Class Codes: G06Q 30/0251 Targeted advertisements
• CPC Class Codes: G06Q 30/0631 Item recommendations

Abstract

A system and method for real-time generation of segments for a computerized decision system provides an event segment list to a decision engine. The event segment list is generated based on consumer affinities determined from consumer data and modeled data on that consumer'"'"'s consumption and geo-location. The consumer affinities are used in a query to obtain current events related to the consumer'"'"'s affinities. The current events are associated with generic events, which together are included in the event segment list, which is delivered to the decision engine and used to make a decision based on the event segment list and specified decision criteria.

Claims

1. A computer-based method for generating a real-time data segment list file for use by a computer-based decision engine to make a real-time decision based on real-time events and consumer-specific real-time data segments in the real-time data segment list file, comprising:

receiving a data segment list file request from the decision engine;

generating a list of consumer primary affinities based on at least one of consumer-declared affinities data, consumer consumption data, and consumer geolocation data;

determining a relevant time window for relevant current events;

generating a list of current event real-time data segments occurring during the relevant time window that relate to the consumer primary affinities, based on current event real-time data segments received from at least one event data source;

retrieving a list of generic event data segments indicative of generic events, having a corresponding predetermined predicted real-time emotional state of the consumer and a corresponding predetermined emotion duration time, associated with each generic event

dynamically associating by the computer each current event real-time data segment for each real-time event with at least one of the predetermined generic event data segments, the predetermined generic event data segments having at least the corresponding predicted real-time emotional state of the consumer and the corresponding emotion duration time;

creating by the computer the real-time data segment list file using results of the associating, such that each real-time event has at least one generic event data segment having the corresponding predicted real-time emotional state of the consumer and the corresponding emotion duration time associated with the generic event;

sending by the computer the real-time data segment list file, including at least the predicted real-time emotional state of the consumer and a corresponding emotion duration time, to the decision engine for use in making an accurate real-time decision regarding content to be delivered to the consumer, the response time from receiving the data segment list file request to sending the data segment list file being less than 500 milliseconds; and

generating by the decision engine, in response to receiving the data segment list file, a recommendation of content to be consumed by the consumer or a selection of an advertisement to be shown to the consumer, the content recommendation or advertisement selection being based at least on the predicted real-time emotional state of the consumer and the corresponding emotion duration time.

2. The method of claim 1, wherein:

the associating each current event segment with a predetermined generic event segment includes applying a set of predetermined rules to the current event segment.

3. The method of claim 1, wherein:

the decision engine is a content recommendation engine and the real-time decision regarding content to be delivered to the consumer is a recommendation of content to be consumed by the consumer.

4. The method of claim 3, wherein:

the content to be recommended is a broadcast of a sporting event.

5. The method of claim 1, wherein:

the decision engine is an advertisement selection engine and the real-time decision regarding content to be delivered to the consumer is a selection of an advertisement to be shown to the consumer based on the real-time data segment list file and specified advertiser decision criteria.

6. The method of claim 1, wherein the generating the consumer primary affinities list comprises:

determining whether consumer-declared affinities are available for the consumer, and if they are, generating a list of the consumer-declared affinities;

determining whether consumer consumption data related to the consumer'"'"'s consumption is available and if it is, calculating at least one consumption score for the consumer based on the consumer consumption data;

if no consumer-declared affinities are available for the consumer, generating at least one estimated consumer affinity based on the at least one consumption score;

determining the consumer'"'"'s geo location;

determining context data related to the consumer'"'"'s context;

determining primary affinities from at least one of consumer declared affinities and estimated consumer affinities by applying the at least one consumption score to the at least one of consumer declared affinities and estimated consumer affinities to break any ties, and if ties remain, applying a geo location model to the affinities to generate the primary affinities list.

7. The method of claim 6, wherein:

the determining context data related to the consumer'"'"'s context includes calculating a content consumption score based on content previously consumed by the consumer that is related to consumer affinities.

8. The method of claim 7, wherein:

the content consumption score includes a recency-frequency score based on recency and frequency data.

9. The method of claim 7, wherein:

the content consumption score includes a recency-frequency score based on recency, frequency, and scale data.

10. The method of claim 1, wherein the generating the consumer primary affinities list is performed before the receiving the request.

11. A computer-based method for generating a real-time data segment list file for use by a computerized decision engine to make a real-time decision based on current sports real-time events and sports-related consumer-specific real-time data segments, comprising:

receiving a sports-related data segment list file request from the decision engine;

generating a list of consumer primary team fan affinities based on at least one of consumer-declared fan affinities data, consumer sports content consumption data, and consumer geolocation data;

determining a relevant time window for relevant current sports events;

generating a list of current sports event real-time data segments occurring during the relevant time window that relate to the consumer primary team fan affinities, based on current sports event real-time data segments received from at least one sports event data source;

retrieving a list of generic sports event data segments indicative of generic sports events, having a corresponding predetermined predicted real-time emotional state of the consumer and a corresponding predetermined emotion duration time, associated with each generic sports event;

dynamically associating by the computer each current sports event real-time data segment for each real-time event with at least one of the predetermined generic sports event data segments, the predetermined generic sports event data segments having at least the corresponding predicted real-time emotional state of the consumer and the corresponding emotion duration time;

creating by the computer the real-time data segment list file using results of the associating, such that each real-time event has at least one generic event data segment having the corresponding predicted real-time emotional state of the consumer and the corresponding emotion duration time associated with the generic event;

sending by the computer the real-time data segment list file, including at least the predicted real-time emotional data of the consumer and a corresponding emotion duration time, to the decision engine for use in making an accurate real-time decision regarding content to be delivered to the consumer, the response time from receiving the data segment list file request to sending the data segment list file being less than 500 milliseconds; and

generating by the decision engine, in response to receiving the data segment list file, a recommendation of content to be consumed by the consumer or a selection of an advertisement to be shown to the consumer, the content recommendation or advertisement selection being based at least on the predicted real-time emotional state of the consumer and the corresponding emotion duration time.

12. The method of claim 11, wherein:

the generating the list of consumer primary team fan affinities includes generating a list of players on a fantasy sports team of the consumer;

the generating the list of current sports event segments that relate to the consumer primary team fan affinities includes generating at least one current sports event segment that relates to the performance of the consumer'"'"'s fantasy sports team.

13. The method of claim 11, wherein the generating the consumer primary team fan affinities list comprises:

determining whether consumer-declared team fan affinities are available for the consumer, and if they are, generating a list of the consumer-declared team fan affinities;

determining whether consumer sports content consumption data related to the consumer'"'"'s consumption of sports content is available and if it is, calculating at least one consumption score for the consumer based on the consumer sports content consumption data;

if no consumer-declared team fan affinities are available for the consumer, generating at least one estimated consumer team fan affinity based on the at least one consumption score;

determining the consumer'"'"'s geo location;

determining content data related to sports content the consumer is consuming; and

determining primary team fan affinities from at least one of consumer declared affinities and estimated consumer team fan affinities by applying the at least one consumption score to the at least one of consumer declared team fan affinities and estimated consumer team fan affinities to break any ties, and if ties remain, applying a geo location model to the team fan affinities to generate the primary team fan affinities list.

14. The method of claim 11, wherein the generating the consumer primary affinities list is performed before the receiving the request.

15. A computer-based method for generating a real-time data segment list file for use by a computer-based decision engine to make a real-time decision based on real-time events and consumer-specific real-time data segments in the real-time data segment list file, comprising:

receiving a data segment list file request from the decision engine;

generating a list of consumer primary affinities based on at least one of consumer-declared affinities data, consumer consumption data, and consumer geolocation data;

determining a relevant time window for relevant current events;

generating a list of current event real-time data segments occurring during the relevant time window that relate to the consumer primary affinities, based on current event real-time data segments received from at least one event data source;

retrieving a list of generic event data segments indicative of generic events, having a corresponding predetermined predicted real-time emotional state of the consumer associated with each generic event;

dynamically associating by the computer each current event real-time data segment for each real-time event with at least one of the predetermined generic event data segments, the predetermined generic event data segments having at least the corresponding predicted real-time emotional state of the consumer;

creating by the computer the real-time data segment list file using results of the associating, such that each real-time event has at least one generic event data segment having the corresponding predicted real-time emotional state of the consumer associated with the generic event;

sending by the computer the real-time data segment list file, including at least the predicted real-time emotional state of the consumer, to the decision engine for use in making an accurate real-time decision regarding content to be delivered to the consumer, the response time from receiving the data segment list file request to sending the data segment list file being less than 500 milliseconds; and

generating by the decision engine, in response to receiving the data segment list file, a recommendation of content to be consumed by the consumer or a selection of an advertisement to be shown to the consumer, the content recommendation or advertisement selection being based at least on the predicted real-time emotional state of the consumer.

16. The method of claim 15, wherein the generating the consumer primary affinities list is performed before the receiving the request.

17. The method of claim 15, wherein the decision engine is a content recommendation engine and the real-time decision regarding content to be delivered to the consumer is a recommendation of content to be consumed by the consumer.

18. The method of claim 15, wherein the decision engine is an advertisement selection engine and the real-time decision regarding content to be delivered to the consumer is a selection of an advertisement to be shown to the consumer based on the real-time data segment list file and specified advertiser decision criteria.