The main .ipynb file that houses the project is titled "musicClassification" and can be found here: https://github.com/nikhilbhave9/Perfect-Pitch/blob/main/musicClassification.ipynb
Serves as final project for CS-1390: Introduction to Machine Learning (Monsoon 2021)
Contributors:
Our project’s goal was to classify a user’s custom inputted Spotify playlist based on the genre of the individual songs. Spotify’s official API does not specify the genre of the song. As such, a predictive ML algorithm to figure out the genre of a song has great value. The three algorithms that we decided to use and compare were SVM, KMC, and Bagging. Our model was trained on the GTZAN dataset. The GTZAN dataset contains 1000 songs, each song of about 30 seconds, 100 of each genre,: blues, classical, country, disco, hiphop, jazz, metal, pop, reggae, rock. The files were collected in 2000-2001 from a variety of sources like individual CDs, radio, microphone recordings, in order to represent a variety of recording conditions. The tracks are all 22050Hz Mono 16-bit audio files in .wav format.