Algorithmic Musical Genre Classification

In this project, I construct a data pipeline which intakes raw .wav files, and then uses machine learning to predict the genre of the track. We first do a frequency-space transformation (similar to the Fourier transform), and then do randomized dimension reduction on the resulting array. Finally, we put the dimension-reduced signal through a naive Bayes classifier, which we train on about 500 sample tracks…

If you are not automatically redirected, please click here