This repository collects the code for the project that Alessandro Pedrani and Matteo Rossi have done for the course Statistical Machine Learning (MAT0043) at the University of Turin, under the guidance of Prof. Silvia Montagna.
- 1 - Features_Extraction.ipynb is the notebook contating the code to extract MFCCs with different number of coefficients and different sampling rates and storing then in a (huge) .txt file.
- 2 - Visualization.ipynb contains some visual exploratory analysis on the data.
- 3 - Classification at Frame-Level using KNN.ipynb, 4 - Classification using GMM.ipynb, 5 - Classification using GMM+frame selection.ipynb and 6 - Classification using a Bayesian Gaussian Mixture.ipynb are the files for the splitting the database and running our models and their extensions.
- 7 - Test a new Song with the best model (for FUN).ipynb is a fun notebook for visualizing exactly how frames of a song are classified by our best model (saved in the file GMM.joblib).