Let's Replay That Song Again
Additional Funding Sources
The project described was supported by the Ronald E. McNair Post-Baccalaureate Achievement Program through the U.S. Department of Education under Award No. P217A170273.
Abstract
Recommender systems have been used in many domains, one of them being music. As a way to better serve users, music platforms like Spotify are known to suggest their user’s songs that they are familiar with. Recent studies have offered insights into how to prioritize songs from users’ collections which hasn’t been played in a while. This has been accomplished by examining interesting links between songs based on their metadata-more specifically textual-based metadata. To further advance knowledge in this area, we propose to explore an effective dimension for recommending user’s songs which have not been replayed in some time. In particular we will explore trends that emerge as a result of generating recommendation informed by textual similarities based on emotions. We will rely on existing experimental frameworks that entail the use the greedy and Hill climbing algorithms as well as Linguistic Inquiry and Word Count (LWIC) for emotion estimation. In this analysis, we expect to reveal emotional connections among songs that can impact the recommendation process and whether emotions influence the prioritization of songs.
Let's Replay That Song Again
Recommender systems have been used in many domains, one of them being music. As a way to better serve users, music platforms like Spotify are known to suggest their user’s songs that they are familiar with. Recent studies have offered insights into how to prioritize songs from users’ collections which hasn’t been played in a while. This has been accomplished by examining interesting links between songs based on their metadata-more specifically textual-based metadata. To further advance knowledge in this area, we propose to explore an effective dimension for recommending user’s songs which have not been replayed in some time. In particular we will explore trends that emerge as a result of generating recommendation informed by textual similarities based on emotions. We will rely on existing experimental frameworks that entail the use the greedy and Hill climbing algorithms as well as Linguistic Inquiry and Word Count (LWIC) for emotion estimation. In this analysis, we expect to reveal emotional connections among songs that can impact the recommendation process and whether emotions influence the prioritization of songs.