<![CDATA[Digital music has grown dramatically in recent years. They offering musics in various type and emotion that are random. Therefore, there is a need to organize the songs, they need to somehow clusterize their files based on a specific characteristic. The purpose of such organization is to enable users to navigate to pieces of music they like, and also to give them advice and recommendation for people or music-related industries. This research proposed a clustering of songs based on their emotion using the Fuzzy c-means algorithm. Audio attributes of valence, energy, loudness, and tempo are used as features that represent the emotions of the song. The cluster of each data is determined based on their membership degree. Cluster validity index is used to evaluate the fitness of partitions produced by clustering algorithms. The algorithm is tested on different amount of data, which is 20%, 40%, 60%, 80%, and 100% data of total 150 songs. The testing result obtained a minimum error value of 0.00000001 (1x10-8). The results showed that the optimal number of clusters that are best to be used in this research is 5. While, the optimal fuzzifier value to be used in this research is 2 with the cluster validity value reaches 0.7 or 70%]]>
Copyrights © 2017
