Play history based detection of similar digital works or work creators

• US 10,275,404 B2
• Filed: 05/20/2013
• Issued: 04/30/2019
• Est. Priority Date: 04/19/2001
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Harbick, Andrew V., Snodgrass, Ryan Justin, Spiegel, Joel R.
• Primary Examiner(s): Kolosowski-Gager, Katherine
• Application Number:
  US13/898,192
• Publication Number:
  US 20130254065A1
• Time in Patent Office: 2,171 Days
• Field of Search: 705 14
• CPC Class Codes: G06F 16/24578 using ranking
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0631 Item recommendations
• CPC Class Codes: H04L 51/02 using automatic reactions or user delegation, e.g. automatic replies or chatbot-generated messages

Abstract

A computer-implemented service recommends digital works (and/or creators of works) to a user based on works currently or previously played or downloaded by the user on a player device or based on playlists stored on the player device. The works may be, for example, music files, video files, electronic books, or other digital content for playing by users. A user may thus obtain personalized recommendations that are based on works obtained from sources (web sites, physical CDs, etc.) that are independent of the recommendations system. In one embodiment, the service identifies pairs of works (and/or work creators) that are similar to each other by virtue of the relatively high frequency with which they co-occur on playlists or within play histories of users. The resulting mappings are used to provide recommendations to users.

Claims

1. Non-transitory computer storage having stored thereon executable code that directs a computing system to implement a process that comprises:

storing play event metadata received over a network from each of a plurality of player devices of each of a plurality of users, said metadata associated with, and including information extracted by the player devices from, digital works played by the users on the player devices, the metadata reported by the player devices to the computing system over the network in response to play events in which digital works are played on the player devices, the metadata further including play event timestamps;

generating session-specific play histories using the received metadata, each session-specific play history identifying, for a particular play session, a set of digital works played during the particular play session, wherein generating the session-specific play histories comprises using the play event timestamps to associate particular digital works with particular play sessions;

calculating a data value representing a degree to which a first digital work and a second digital work are related to each other based, at least in part, on a measure of how frequently the first and second digital works co-occur in the session-specific play histories, wherein calculating the data value comprises determining a number of said play sessions in which both the first digital work and the second digital work were played;

storing the data value in computer storage;

detecting a play event in which the first digital work is selected for playing on a player device; and

in response to the play event, and based on the data value, automatically transmitting to the player device over the network at least one of (1) the second digital work, and (2) a user interface display that recommends the second digital work.

2. The non-transitory computer storage of claim 1, wherein the process comprises dividing a user'"'"'s play history into multiple sessions, and analyzing each of the multiple sessions separately.

3. The non-transitory computer storage of claim 1, wherein the process further comprises using the data value to provide digital work recommendations to users.

4. The non-transitory computer storage of claim 3, wherein at least some of said recommendations are provided to users within an interactive electronic catalog that provides functionality for users to purchase digital works.

5. The non-transitory computer storage of claim 1, wherein the metadata comprises metadata read from the digital works by the player devices.

6. The non-transitory computer storage of claim 5, wherein the metadata read from the digital works comprises album/artist/track information encoded in the digital works.

7. The non-transitory computer storage of claim 1, wherein the metadata comprises data obtained by the player devices from user-generated playlists.

8. The non-transitory computer storage of claim 1, wherein the player devices include general purpose computing devices that run a player application.

9. The non-transitory computer storage of claim 1, wherein the process further comprises using the session-specific play histories to identify digital work creators that are related to each other.

10. The non-transitory computer storage of claim 1, in combination with said computing system, wherein the computing system is programmed via said executable code to implement said process.

11. Non-transitory computer storage having stored thereon executable code that directs a computing system to implement a process that comprises:

receiving play event metadata over a network from each of a plurality of player devices of each of a plurality of users, said metadata descriptive of, and including information extracted by the player devices from, digital works played by the users on said player devices, the metadata reported by the player devices to the computing system over the network in response to play events in which digital works are played on the player devices, the metadata further including play event timestamps;

generating session-specific play histories using the received metadata, each session-specific play history specifying, for a particular play session, a set of digital works played during the particular play session, wherein generating the session-specific play histories comprises using the play event timestamps to associate particular digital works with particular play sessions;

calculating a data value representing a degree to which a first digital work creator and a second digital work creator are related to each other based, at least in part, on a measure of how frequently digital works of the first and second digital work creators co-occur in the session-specific play histories, wherein calculating the data value comprises determining a number of said play sessions in which a digital work of the first digital work creator and a digital work of the second digital work creator were both played;

storing the data value in computer storage;

detecting a play event in which a digital work of the first digital work creator is played on a player device; and

in response to the play event, and based on the data value, transmitting to the player device a user interface display that recommends at least one of (1) the second digital work creator, or (2) a particular digital work of the second digital work creator.

12. The non-transitory computer storage of claim 11, wherein the process comprises dividing a user'"'"'s play history into multiple sessions, and analyzing each of the multiple sessions separately.

13. The non-transitory computer storage of claim 11, wherein the metadata comprises metadata read from the digital works by the player devices.

14. The non-transitory computer storage of claim 11, wherein the metadata comprises data obtained by the player devices from user-generated playlists.

15. The non-transitory computer storage of claim 11, in combination with said computing system, wherein the computing system is programmed with said executable code to implement said process.

16. A system for detecting relationships, comprising:

a metadata reporting component that runs on player devices of users, said metadata reporting component comprising executable code that directs the player devices to monitor digital work play events on the player devices and to report, on a network, metadata associated with said play events, including metadata extracted from digital files of played digital works, the metadata including play event timestamps; and

a server system programmed to use the metadata reported by the player devices to maintain session-specific digital work play histories, each of which corresponds to a specific play session, said server system additionally programmed to detect relationships between particular digital works, and/or relationships between particular digital work creators, at least partly by analyzing the session-specific play histories, wherein analyzing the session-specific play histories comprises counting numbers of sessions in which two particular digital works were both played;

wherein the server system is programmed to use the play event timestamps to associate particular play events with particular play sessions;

wherein the server system in responsive to a play event in which a first digital work of a first digital work creator is played on a player device by using the detected relationships to select at least one of (1) a second digital work to recommend on the player device, and (2) a second digital work creator to recommend on the player device.

17. The system of claim 16, wherein the metadata additionally includes digital work metadata obtained from user-generated playlists.

18. The system of claim 16, wherein the server system is programmed to detect relationships between particular digital work creators.

19. The system of claim 16, wherein the server system is additionally programmed to use the detected relationships to generate personalized recommendations for users.

20. The system of claim 16, wherein the metadata extracted from digital files of played digital works comprises ID3 tag data.

21. The system of claim 16, wherein the metadata extracted from digital files of played digital works comprises album/artist/track information.

22. The system of claim 21, wherein the server system is additionally programmed to execute an algorithm that attempts to match the album/artist/track information extracted from a digital work to a work identifier.