Music recommendation using emotional allocation modeling
An exemplary method includes defining a vocabulary for emotions; extracting descriptions for songs; generating distributions for the songs in an emotion space based at least in part on the vocabulary and the extracted descriptions; extracting salient words from a document; generating a distribution for the document in an emotion space based at least in part on the vocabulary and the extracted salient words; and matching the distribution for the document to one or more of the distributions for the songs. Various other exemplary methods, devices, systems, etc., are also disclosed.
1. A method, implemented at least in part by a computing device, comprising:
defining a vocabulary for emotions;
extracting descriptions for one or more songs from web-based information that describes the one or more songs;
generating one or more first distributions for the one or more songs in an emotion space based at least in part on the vocabulary and the extracted descriptions;
extracting one or more salient words from a document;
generating a second distribution for the document in the emotion space based at least in part on the vocabulary and the extracted salient words; and
at the computing device, comparing the second distribution for the document to at least one of the first distributions for the one or more songs to provide at least one song recommendation corresponding to the second distribution for the document in the emotion space.
2. The method of claim 1 wherein the emotion space includes at least a minimum number of emotions to allow for mixed emotion mapping.
3. The method of claim 1 wherein the comparing comprises matching the first and second distributions based at least in part on Kullback-Leibler divergence.
4. The method of claim 1 wherein the comparing includes matching one or more closest document-song pairs.
5. The method of claim 1 wherein extracting descriptions for songs includes searching the Internet for the web-based information.
6. The method of claim 5 wherein searching the Internet is based at least in part on song titles for the one or more songs.
7. The method of claim 1 wherein the emotion space includes more than 10 emotions.
8. The method of claim 1 further comprising providing via a network an identification of the one or more songs that most closely matched the second distribution for the document.
9. The method of claim 8 further comprising providing information for purchasing the one or more songs.
10. The method of claim 1 further comprising comparing the one or more songs to a personal collection of songs.
11. The method of claim 1 wherein the document comprises a web document.
12. The method of claim 11 wherein the web document comprises a weblog.
13. The method of claim 1 wherein the one or more songs comprise emotional characteristics that match emotional characteristics of the document.
14. One or more computer-readable storage devices comprising computer-executable instructions which, when executed by at least one processor, instruct the at least one processor to perform actions comprising:
by the at least one processor, providing an emotional allocation model to generate distributions for documents and distributions for songs in a common emotional space, the distributions for songs based on web-based information associated with the songs;
by the at least one processor, employing a Kullback-Leibler divergence algorithm to compare the generated distributions for documents and the generated distributions for songs; and
by the at least one processor, providing at least one song recommendation corresponding to the distributions for the documents in the common emotional space.
15. The one or more computer-readable storage devices of claim 14 , wherein the actions further comprise:
extracting salient words from the documents; and
generating the distributions for the documents based on the extracted salient words.
16. The one or more computer-readable storage devices of claim 14 , wherein the actions further comprise extracting descriptions for the songs from the web-based information.
17. A server comprising the one or more computer-readable storage devices of claim 14 .
18. A computing device comprising:
a network interface;
one or more processors;
a memory; and
instructions stored in the memory and executed by the one or more processors to instruct the one or more processors to:
generate distributions for documents and distributions for songs in a common emotional space, wherein the distributions for the songs are based at least in part on web-based information associated with songs;
match a generated distribution for a document to at least one generated distribution for a song; and
output via the network interface information about a song that has a distribution that matches a distribution of a document.
19. The computing device of claim 18 , wherein the instructions further instruct the one or more processors to receive a document via the network interface.
20. The computing device of claim 18 , wherein the instructions further instruct the one or more processors to distribute revenue from sales of the song that has a distribution that matches a distribution of a document.