System and method for selecting media content
Methods, systems, and computer programs for generating a playlist of media content items without explicit content. A vector space is created that represents explicit and non-explicit tracks in the same playlists created by other users and then tracks are filtered based on cosine distance between the “seed tracks” and all the tracks in the aforementioned playlist. The explicit tracks are filtered out, and tracks are sorted based on the affinity of the user to the artist.
1. Determining a clean media content item to play based on a seed media content item that is selected or currently playing, wherein the seed media content item includes explicit content, by:
receiving the seed media content item at an information storage and media distribution platform from a database of a plurality of media content items;
creating a vector space comprising the seed media content item and a plurality of related media content items, the related media content items including explicit media content items, each explicit media content item including explicit content, and a plurality of clean media content items, each clean media content item not including explicit content, wherein none of the plurality of clean media content items is the seed media content item with the explicit content in the seed media content item removed or altered;
calculating a similarity score between the seed media content item and the plurality of related media content items;
removing explicit media content items from the plurality of related media content items;
determining an affinity score for each of the plurality of clean media content items; and
determining the clean media content item to play based on the affinity score and similarity score for each of the plurality of clean media content items; and playing the clean media content item from the plurality of clean media content items.
2. The method according to claim 1 , wherein the similarity score is calculated using a cosine similarity function.
3. The method according to claim 1 , wherein the affinity score is calculated by comparing the plurality of related media content items to a user's listening history.
4. The method according to claim 1 , wherein the plurality of related media content items are taken from playlists that comprise at least 50% explicit media content items.
5. The method according to claim 1 , further comprising generating a list of media content items from the plurality of clean media content items in an order based on the affinity score.
6. The method according to claim 1 , wherein the seed media content item is an audio track and the plurality of related media content items are selected from a plurality of audio tracks.
7. A system comprising a non-transitory memory and a processing device coupled to the memory and constructed to execute the process steps stored in the memory, wherein the stored process steps include steps which when executed by the processor cause the processor to perform the steps of: determining a clean media content item to play based on a seed media content item that is selected or currently playing, wherein the seed media content item includes explicit content, by:
receiving the seed media content item at an information storage and retrieval platform from a database of a plurality of media content item;
creating a vector space comprising the seed media content item and a plurality of related media content items, the related media content items including explicit media content items, each explicit media content item including explicit content, and a plurality of clean media content items, each clean media content item not including explicit content, wherein none of the plurality of clean media content items is the seed media content item with the explicit content in the seed media content item removed or altered;
calculating a similarity score between the seed media content item and the plurality of related media content items;
removing explicit media content items from the plurality of related media content items;
determining an affinity score for each of the plurality of clean media content items; and
determining the clean media content item to play based on the affinity score and the similarity score for each of the plurality of clean media content items; and playing the clean media content item from the plurality of clean media content items based.
8. The system according to claim 7 , wherein the similarity score is calculated using a cosine similarity function.
9. The system according to claim 7 , wherein the affinity score is calculated by comparing the plurality of related media content items to a user's listening history.
10. The system according to claim 7 , wherein the plurality of related media content items are taken from playlists that comprise at least 50% explicit media content items.
11. The system according to claim 7 , further comprising generating a list of media content items from the plurality of clean media content items in an order based on the affinity score.
12. The system according to claim 7 , wherein the seed media content item is an audio track and the plurality of related media content items are selected from a plurality of audio tracks.
13. The method of claim 1 , wherein the seed media content item is a currently playing media content item.
14. The system of claim 7 , wherein the seed media content item is a currently playing media content item.
15. The method according to claim 1 , wherein calculating the similarity score between the seed media content item and the plurality of related media content items comprises determining if the plurality of related media content items are in one or more playlists that include the seed media content item.
16. The system according to claim 7 , wherein calculating the similarity score between the seed media content item and the plurality of related media content items comprises determining if the plurality of related media content items are in one or more playlists that include the seed media content item.
17. The method according to claim 1 , wherein calculating the similarity score between the seed media content item and the plurality of clean media content items comprises comparing the plurality of related media content items to a user's taste profile.
18. The system according to claim 7 , wherein calculating the similarity score between the seed media content item and the plurality of clean media content items comprises comparing the plurality of related media content items to a user's taste profile.