With the dominance of digital music streaming, music genre classification is an increasingly important way to develop algorithms for suggesting new songs to users. Without a way to analyze the abstract aspects of each song such as mood and tone that people use to determine genre, these algorithms must be able to make accurate predictions based only on the quantitative characteristics of the audio itself.
Hence, we aim to answer the following research question:
Can we predict the genre of a song based on its audio spectrograph information?
In particular, we will focus on the classical, jazz, pop, and rock music genres.