Publication Date

Spring 2019

Degree Type

Master's Project

Degree Name

Master of Science (MS)


Computer Science

First Advisor

Teng Moh

Second Advisor

Chris Pollett

Third Advisor

Kevin Smith


music recommendation, music genre classification, Hierarchical Attention Networks


With the advent of digitized music, many online streaming companies such as Spotify have capitalized on a listener’s need for a common stream platform. An essential component of such a platform is the recommender systems that suggest to the constituent user base, related tracks, albums and artists. In order to sustain such a recommender system, labeling data to indicate which genre it belongs to is essential. Most recent academic publications that deal with music genre classification focus on the use of deep neural networks developed and applied within the music genre classification domain. This thesis attempts to use some of the highly sophisticated techniques, such as Hierarchical Attention Networks that exist within the text classification domain in order to classify tracks of different genres. In order to do this, the music is first separated into different tracks (drums, vocals, bass and accompaniment) and converted into symbolic text data. Due to the sophistication of the distributed machine learning system (over five computers, each possessing a graphical processing units greater than a GTX 1070) present in this thesis, it is capable of classifying contemporary genres with an impressive peak accuracy of over 93%, when comparing the results with that of competing classifiers. It is also argued that through the use text classification, the ex- pert domain knowledge which musicians and people involved with musicological techniques, can be attracted to improving reccomender systems within the music information retrieval research domain.