Music recommendation system based on context using case-based reasoning and self organizing map

Abstract

Choosing music from a music library in a computer that consists hundreds or thousands of music can be time consuming. An effective music recommendation can decrease listener’s effort in choosing music that will be listened. Music recommendation is not only can be obtained based on genre or audio similarity, because listener’s music choices are also influenced by the listener’s context (mood, occasion, part of day, date, weather, region, month, and weekday). This research used Case-Based Reasoning (CBR) for determining music recommendation based on listener’s context data and also Self Organizing Map (SOM) which is used as an indexing method in CBR. Inputs given by user to the system are user’s occasion and mood desired by user. The system output is a playlist consists of music that suitable with user’s context and desired mood.Precision value produced by the system that used SOM as an indexing method is 0.867. It showed that the recommendation produced by the system is 86.7% relevant according to the user.