Deep Learning Techniques for Music Genre Classification and Feature Importance

Abstract

<p>Utilized GTZAN dataset which consisted of features extracted at 3-second and .3-second intervals and includes features such as mel-frequency cepstrum, tempo, and harmony. Additionally, I used the raw audio signals and developed a 1D CNN for the classification of music genres. Experimented between different default and complex machine learning models for the music genre classification task on the GTZAN dataset as a whole.</p>