Personalized generation of watch list of shows in a video delivery system

Patent metadata

Field	Value
Publication number	US-9560399-B2
Application number	US-201514738376-A
Country	US
Kind code	B2
Filing date	Jun 12, 2015
Priority date	Jun 13, 2014
Publication date	Jan 31, 2017
Grant date	Jan 31, 2017

Abstract

Particular embodiments provide a watch list of shows to users. The watch list is personalized for each user. Also, the watch list is dynamically organized to predict an order the user will want to watch the shows. Particular embodiments analyze historical user behavior with respect to the timing for recurring releases of the episodes for shows to determine the order of the shows in the watch list. The watch list is organized in a way that a user may select a “watch all” button where unseen episodes for the shows in the watch list are all played to the user in an order that is predicted to be the order in which the user would want to watch the shows. Providing the watch all button makes it important to predict the order of the shows accurately.

First claim

What is claimed is:

1. A method comprising:
storing, by a computing device, a watch history of shows for a user at a video delivery service that provides a service to send videos to a plurality of users, wherein the shows release recurring episodes in a season on the video delivery service;
training, by the computing device, a machine learning predictor to determine selection probabilities for shows using the watch history of shows, the selection probability indicating a prediction the user will select the show to watch;
receiving, by the computing device, a request for an episode of a show from a client device being used by the user;
updating, by the computing device, the watch history of shows with an indication the episode of the show was watched;
determining, by the computing device, a list of eligible shows for the watch list based on the updated watch history of shows for the user;
receiving, by the computing device, user specific features for the list of eligible shows based on the updated user's watch history, wherein the updated user's watch history is used to determine unseen released episodes for at least a portion of the list of eligible shows;
inputting, by the computing device, the user specific features and the list of eligible shows into the machine learning predictor to determine selection probabilities for each of the list of eligible shows, the selection probability indicating the prediction the user will select the show to watch and affected by a number of unseen released episodes for the show when the show has unseen released episodes; and
dynamically updating, by the computing device, the watch list of shows for the user based on the selection probabilities of the list of eligible shows, wherein the watch list of shows orders the shows in an order in which the selection probability predicts that the user will select one or more of the list of eligible shows to watch.

2. The method of claim 1 , further comprising:
outputting a watch all button configured to play the watch list of shows in the order without further user input; and
upon receiving an input selecting the watch all button, causing unseen released episodes of the watch list of shows to be sent in the order without further user input.

3. The method of claim 2 , wherein all unseen released episodes for a first show in the order are sent before unseen released episodes of a next show in the order are sent.

4. The method of claim 1 , further comprising:
categorizing the at least a portion of the list of eligible shows into a plurality of categories based on a current status of unseen released episodes for shows in the list of eligible shows.

5. The method of claim 4 , wherein the plurality of categories comprise:
a first category in which one or less unseen released episode has not been watched in a show that is releasing current episodes in a time period,
a second category in which two or more unseen released episodes have not been watched in a show that is releasing current episodes in the time period, and
a third category in which a show is not releasing current episodes in the time period.

6. The method of claim 4 , further comprising:
determining an effective time for each show in the list of eligible shows, the effective time based on the category in the plurality of categories for each show.

7. The method of claim 6 , further comprising:
pre-ordering the list of eligible shows by the effective time in a pre-order; and
grouping shows in the pre-order based on respective categories in the plurality of categories of the shows.

8. The method of claim 7 , further comprising:
using the machine learning predictor to order shows within each group.

9. The method of claim 4 , further comprising:
using the machine learning predictor to order shows in the watch list using the effective time and a category in the plurality of categories for each show as input to the machine learning predictor.

10. A method comprising:
receiving, by a computing device, a watch history of shows for a user at a video delivery service that provides a service to send videos to a plurality of users, wherein the shows release recurring episodes in a season on the video delivery service;
determining, by the computing device, a list of eligible shows for a watch list based on the watch history of shows for the user;
categorizing, by the computing device, the list of eligible shows into a plurality of categories based on a current status of a number of unseen released episodes for shows in the list of eligible shows;
determining, by the computing device, an effective time for each show in the list of shows, the effective time based on the category for each show;
grouping, by the computing device, the list of shows into a plurality of groups based on the categories and the effective time for each show;
receiving, by the computing device, user specific features for the list of eligible shows based on the user's watch history;
for each group in the plurality of groups, inputting, by the computing device, the user specific features and the shows in a group into a machine learning predictor to determine a first order the shows in the group; and
outputting, by the computing device, the watch list of shows in a second order based on the first order within the groups.

11. The method of claim 10 , further comprising:
outputting a watch all button configured to play the watch list of shows in the second order without further user input; and
upon receiving an input selecting the watch all button, causing unseen released episodes of the watch list of shows to be sent in the watch list order without further user input.

12. The method of claim 10 , wherein the plurality of categories comprise:
a first category in which one or less next unseen released episode has not been watched in a show that is releasing current episodes in a time period,
a second category in which two or more next unseen released episodes have not been watched in a show that is releasing current episodes in the time period, and
a third category in which a show is not releasing current episodes in the time period.

13. The method of claim 12 , further comprising:
ordering the list of eligible shows into a third order based on the effective time before performing the grouping,
wherein the grouping is performed to group consecutive shows with a category considered to be of equal value.

14. The method of claim 13 , wherein:
consecutive shows in the first category are grouped in first groups, and
consecutive shows in the second category and third category are grouped in second groups.

15. The method of claim 14 , wherein:
the machine learning predictor orders shows within each of the first groups and second groups.

16. The method of claim 10 , wherein:
the effective time for the first category is a release time on the next unseen released episode, and
the effective time for the second category and the third category is a watch time of a newest episode.

17. An apparatus comprising:
one or more computer processors; and
a non-transitory computer-readable storage medium comprising instructions, that when executed, control the one or more computer processors to be configured for:
storing a watch history of shows for a user at a video delivery service that provides a service to send videos to a plurality of users, wherein the shows release recurring episodes in a season on the video delivery service;
training a machine learning predictor to determine selection probabilities for shows using the watch history of shows, t

Assignees

• Hulu Llc

Inventors

• Kaya Ilke
• Elston Devin
• Tang Bangsheng
• Yao Jinyu
• Su Heng
• Liu Mingkui
• Kolasinski Jordan

Classifications

• H04N21/44222

  Analytics of user selections, e.g. selection of programmes or purchase activity (monitoring of user selections in data processing systems G06F11/34; arrangements for monitoring the user's behaviour or opinions in broadcast systems H04H60/33) · CPC title

• H04N21/4663

  involving probabilistic networks, e.g. Bayesian networks · CPC title

• H04N21/251 Primary

  Learning process for intelligent management, e.g. learning user preferences for recommending movies (details of learning user preferences for the retrieval of video data in a video database G06F16/739; computer systems using learning methods G06N3/08) · CPC title

• H04N21/6582

  Data stored in the client, e.g. viewing habits, hardware capabilities, credit card number (arrangements where receivers interact with the broadcast H04H20/38) · CPC title

• H04N21/4826

  using recommendation lists, e.g. of programmes or channels sorted out according to their score · CPC title