Music Mood Prediction Based on Spotify’s Audio Features Using Logistic Regression
Marvin Ray Dalida, Lyah Bianca Aquino, William Cris Hod, Rachelle Ann Agapor, Shekinah Lor Huyo-a, Gabriel Avelino Sampedro · 2022
Music influences our mood. Individuals have experienced music personally where their emotions become involved, allowing the tempo or lyrics of the music to impact them. Music not only provides entertainment but also helps boost overall well-being. Over the years, music streaming platforms have become popular for the way music is delivered and music queues have been tailored for the listener. In these applications, machine learning has been for music recommendation. In this paper, an innovative approach for modeling a track’s mood based on audio components from the Spotify application program interface (API) available on the Spotify PH market will be explored. This paper will focus on the performance evaluation of the the application of logistic regression in predicting the mood of a song. The validation of the model uses the stratified k-fold cross-validation and evaluation confusion matrix for analyzing the model. The primary expected outcome of the system is to predict the song’s mood based on the 12 features of Spotify audio components. The model will be the basis of future studies to identify the best factor that affects the track’s mood based on its audio features.