Music Recommendation System Based on Context and EEG Data

An effective music recommendation can reduce the effort given by a music listener inchoosing a piece of music to be heard. Music recommendations can not only be obtained bysimilar genre or audio similarity, because the music chosen by the music listeners can bedifferent in different contexts. In this research, a case base for music recommendation systemwas developed based on the music listeners’ context in order to make it easier to choosemusic that suits the current situation. The context used in this study is the personal context ofthe music listeners consisting of age, gender, emotional states, favorite activities, and musicalpreferences by genre. An emotional state often called a mood will be detected usingbrainwave sensors. Brainwaves or Electro Encephalogram (EEG) will be classified based onthe emotional state of the listeners and the Case based reasoning (CBR) method used todetermine music recommendations based on the context of the listener and EEG data. Thesystem tested to 20 participants and obtained music recommendation accuracy value of64.5%.An effective music recommendation can reduce the effort given by a music listener inchoosing a piece of music to be heard. Music recommendations can not only be obtained bysimilar genre or audio similarity, because the music chosen by the music listeners can bedifferent in different contexts. In this research, a case base for music recommendation systemwas developed based on the music listeners’ context in order to make it easier to choosemusic that suits the current situation. The context used in this study is the personal context ofthe music listeners consisting of age, gender, emotional states, favorite activities, and musicalpreferences by genre. An emotional state often called a mood will be detected usingbrainwave sensors. Brainwaves or Electro Encephalogram (EEG) will be classified based onthe emotional state of the listeners and the Case based reasoning (CBR) method used todetermine music recommendations based on the context of the listener and EEG data. Thesystem tested to 20 participants and obtained music recommendation accuracy value of64.5%.