Music Recommendation System Using Content-based Filtering Method with Euclidean Distance Algorithm

Abstract

For some people listening to music can soothe the heart and soul, especially if people listen to music that suits their favourite. There are problems faced by listeners when choosing the music they want that fits the user's mood. In this study, the authors used the content-based filtering method to find music recommendations and used the mood model of James Russel's Circumplex as a reference for user mood. With input from the user's mood, it will be processed using the Euclidean distance algorithm to calculate the distance from the item to other items and get the top-n recommendation with the smallest distance. The output of this study is the top 10 songs with the smallest distance according to user mood input. Furthermore, an evaluation of the recommendations using the NDCG technique was carried out, and an average NDCG value of 0.95235 was obtained closer to 1. It means the music recommendations given are very relevant.