SISTEM REKOMENDASI MUSIK BERDASARKAN DATA KONTEKS PADA LISTENING HISTORY MUSIK DAN KETERKAITAN ARTIS INDONESIA

Abstract

A large number of digital music circulates online today. It makes music listeners confused to choose which music is suitable to listen to in certain circumstances or contexts, for example certain time, weather, activity, and desired mood. Playlist creation can make it easy for music listeners to collect their favorite music for a particular context, but creating playlists is time consuming and of course a lot of playlists will have to be created to accommodate all combinations of contexts. In this study, an automated music recommendation system was built using context data consisting of time, weather, activities, and desired mood which were also adjusted for the listener's age, gender, and favorite artist. The method used is Case-Based Reasoning (CBR), using listeners' listening history data as a knowledge base and the artist relatedness of Indonesian artists to improve solutions at the revision stage. Output of this system is in the form of music playlist presented in a website. The overall precision average for music recommendations is 0.78.