Aplikasi Rekomendasi Musik Berdasarkan Klasifikasi Genre Menggunakan Convolutional Neural Networks

Abstract

Convolutional Neural Network can be implemented on various deep learning approach for scientific or problem solving real world case. This research discusess implementing Convolutional Neural Network for music genre classification. Researcher will classified 8 genres music on this research. Music will be converted first to MFCC so its features can be extracted. Researcher use GTZAN Dataset as the base line for Convolutional Neural Network to learn each feature on a bunch of genres. Validation Accruacy with GTZAN dataset is 79% and Real world classification resulting in 60%. The classification progam can be used on android devices. This research also serve music suggestions based the result of predicted genre.