This project explores the use of machine learning to detect anomalies in the musical style of the band Primus. Inspired by their unconventional sound and unique approach to music, this research aims to analyze their genre and identify elements that make their music stand out from traditional classifications.
As a HUGE fan of the prog-metal band TOOL from LA, one day I decided to check out Primus's music and ended up loving it. While I wouldn't describe myself as a die-hard fan of Primus, I find some of their songs incredible. I was also impressed by this part of their Wikipedia page:
... The music of Primus has been described as "thrash-funk meets Don Knotts, Jr." and "the Freak Brothers set to music". The Daily Freeman described the band's style as "a blend of funk metal and experimental rock". The A.V. Club described the band's music as "absurdist funk-rock". Primus have also been described as "prog rock" or "prog metal". AllMusic places Primus within the first wave of alternative metal bands, saying that they fused heavy metal music with progressive rock. Entertainment Weekly classified the band's performance as "prog-rock self-indulgence"...
For context, some of my favorite artists include Iron Maiden, TOOL, Bruce Dickinson, Pink Floyd, Radiohead, Red Hot Chili Peppers, Deftones, Porcupine Tree, Nine Inch Nails, Deep Purple, Metallica, Rush, and others.
Recently, I started learning the basics of machine learning and wanted to apply my new knowledge to a field I’m passionate about.
- By analyzing their discography, I aim to declare an anomaly in music genre classification or find the genre that best "suits" their style.
- This project is not about creating a perfect model but about exploring, making mistakes, and improving along the way.
Important Note
The classification of music genres is subjective! There are no strict rules that define clear borders between two pieces of music from different genres. However, we have some basic divisions created by fans and listeners, such as rock, metal, hip hop, rap, etc.
For research purposes, I will use classifications from Wikipedia as well as my own experience as a music enthusiast. Over the past six years, I’ve recorded details of more than 300 albums in my personal Excel table — and I’m still adding to it.