Media content item recommendation system

Patent metadata

Field	Value
Publication number	US-11086936-B2
Application number	US-201715858377-A
Country	US
Kind code	B2
Filing date	Dec 29, 2017
Priority date	Dec 30, 2016
Publication date	Aug 10, 2021
Grant date	Aug 10, 2021

Abstract

A media content item recommendation system recommends media content items based on one or more attributes of a seed playlist. The recommended media content items can be determined from a plurality of existing playlists that have been created over a period of time. Such existing playlists can be selected based on similarity to the seed playlist.

First claim

What is claimed is:

1. A method for recommending media content items for a playlist, the method comprising:
identifying, using at least one computing device, a plurality of preliminary candidate playlists from a playlist database;
receiving, using the at least one computing device, a title of the playlist;
comparing, using the at least one computing device, the title of the playlist with at least one attribute of each of the plurality of preliminary candidate playlists;
calculating, using the at least one computing device, similarity scores for the plurality of preliminary candidate playlists, each of the similarity scores calculated based on the comparison between the title of the playlist with the at least one attribute of each of the plurality of preliminary candidate playlists;
selecting a plurality of candidate playlists based on the similarity score from the plurality of preliminary candidate playlists;
identifying, using the at least one computing device, one or more media content items from the plurality of candidate playlists;
determining, using the at least one computing device, a frequency that the one or more media content items appear in the plurality of preliminary candidate playlists and a frequency that the one or more media content items appear in the plurality of candidate playlists; and
adding, using the at least one computing device, at least one media content item that appears infrequently in the plurality of preliminary candidate playlists and frequently in the plurality of candidate playlists to the playlist.

2. The method of claim 1 , further comprising receiving user input to create the playlist including at least the title of the playlist.

3. The method of claim 1 , wherein the playlist is an empty playlist.

4. The method of claim 1 , wherein the playlist includes one or more media content item identifiers.

5. The method of claim 4 , wherein attributes of the plurality of preliminary candidate playlists include at least one of a title and an artist identification of media content items contained in the plurality of preliminary candidate playlists.

6. The method of claim 5 , wherein comparing includes:
comparing the titles of media content items of the plurality of preliminary candidate playlists with the title of the playlist.

7. The method of claim 5 , wherein comparing includes:
comparing the artist identifications of the media content items of the plurality of preliminary candidate playlists with the title of the playlist.

8. The method of claim 1 , further comprising:
obtaining popularity scores of the one or more media content items included in the plurality of candidate playlists;
calculating an average popularity score for each of the plurality of candidate playlists, the average popularity score being an average of the popularity scores of the one of more media content items;
employing an aggregation engine suitable for the average popularity score for each of the plurality of candidate playlists;
determining at least one popular media content item from the plurality of candidate playlists based on the aggregation engine; and
adding the at least one popular media content item to the playlist.

9. The method of claim 8 , wherein the aggregation engine is selected from the group consisting of a term frequency-inverse document frequency algorithm, the Mutual Information algorithm, and the Normalized Google Distance.

10. The method of claim 8 , wherein determining the at least one popular media content items further includes:
determining the at least one popular media content items by:
eliminating duplicate media content items from a plurality of candidate media content items, the duplicate media content items already existing in the playlist;
separating media content items associated with an identical artist by a predetermined distance; and
removing user-ignored media content items from the plurality of candidate media content items, the user-ignored media content items including media content items that have been recommended and ignored by a user.

11. The method of claim 1 , further comprising:
automatically updating the playlist with the one or more recommended media content items.

12. The method of claim 1 , wherein identifying the plurality of preliminary candidate playlists includes selecting the plurality of preliminary candidate playlists from the playlist database based on at least one of user activity, playback frequency, number of media content items contained, number of associated artists, and a ratio between the number of media content items and the number of artists.

13. A method of recommending media content items for a seed playlist, the method comprising:
generating a user interface on a computing device, the user interface including a playlist section and a media recommendation section;
receiving a title of the seed playlist through the user interface;
displaying the title of the seed playlist in the playlist section of the user interface;
displaying a list of recommended media content items in the media recommendation section of the user interface, the list of recommended media content items determined by:
identifying a plurality of preliminary candidate playlists from a playlist database;
comparing the title of the seed playlist with at least one attribute of each of the plurality of preliminary candidate playlists;
calculating similarity scores for the plurality of preliminary candidate playlists, each of the similarity scores calculated based on the comparison between the title of the seed playlist with the at least one attribute of each of the plurality of preliminary candidate playlists;
selecting a plurality of candidate playlists based on the similarity score from the plurality of preliminary candidate playlists;
identifying one or more media content items from the plurality of candidate playlists; and
determining a frequency that the one or more media content items appears in the plurality of preliminary candidate playlists and a frequency that the one or more media content items appear in the plurality of candidate playlists; and

enabling a user to add at least one of the media content items to the seed playlist that appears infrequently in the plurality of preliminary candidate playlists and frequently in the plurality of candidate playlists.

14. The method of claim 13 , wherein the recommended media content items have a theme similar to the at least one attribute of the seed playlist.

15. The method of claim 13 , further comprising:
enabling the user to preview the recommended media content items before adding the recommended media content items to the seed playlist.

16. A system for recommending media content items for a seed playlist, the system comprising:
a processing device;
a playlist database storing a plurality of playlists; and
a computer readable data storage medium storing software instructions that, when executed by the processing device, cause the system to:
receive a title of the seed playlist;
identify a plurality of preliminary candidate playlists from the plurality of playlists stored in the playlist database;
compare the title of the seed playlist with an attribute of each of the plurality of preliminary candidate playlists, the attribute of the seed playlist including a description of the seed playlist;
calculate similarity scores for the plurality of preliminary candidate playlists, each of the similarity scores calculated based on the comparison between the title of the seed playlist and the attribute of each of the plurality of preliminary candidate playlists;
s

Assignees

• Spotify Ab

Inventors

• Lamere Paul
• Marmor Jonathan
• Ackerman Michelle
• Lindstrand Per
• Söderlund Oscar
• Johansson Mattias
• Waller Jacob Colin
• Carvalho Felipe Oliveira
• Westraeus Camilla
• Öhman Michael

Classifications

• G06F16/435 Primary

  Filtering based on additional data, e.g. user or group profiles · CPC title

• G06F16/4387

  by the use of playlists · CPC title

• G06F16/90324 Primary

  using system suggestions · CPC title