Moelleux—Music Recommendation System

Abstract

AbstractWith continuous expansion of music resources there was a need for an increase in the rate of searching similar artists which previously consumed a lot of effort as well as time. An automated software was much needed for this which helped in almost instantaneous song recommendation based on similar artists as well as top songs. This article presents a music recommendation using both popularity (Collaborative filtering) and similarity-based (Content-based filtering) approach. In this music recommendation system, uses a well-known filtering method, i.e. content based which will recommend the based on their previously listened and searched songs. Songs will be recommended based on genre, artists and also uses collaborative filtering which will recommend the top ten popular songs from the new releases. The log file contains users’ history. To evaluate the music recommendation system, the million song data set (MSD) was used in the system. To show the working of the music recommendation system, a Web App is developed.KeywordsMusic recommendationContent based filteringCollaborative filteringPopularity-based filteringSimilarity-based filteringTop songs recommendationHistory-based recommendation