Network tour of data sciences project
For this project we use the Free Music Archive (FMA) dataset. it is a free and open library. The goal of this project is to recognize music genre from a track. We are working with music from the genres "Rock","Hip-hop","Experimental" and "Electronic". Do we observe distinct features for every genre and can we identify them? To answer those questions, we are building a similarity graph between music tracks and training Machine Learning model in order to be able to predict their genre and classify them.
https://www.overleaf.com/read/hhmjtyryqgtn
Notebook containing the Extra Trees Classifier model, to run it you need to produce the "data.pkl" file from the "full_data_creation.ipynb" notebook and put it in the folder.
Notebook containig all the other models tested using the 4 genres: "Rock","Hip-hop","Experimental" and "Electronic". To run it you need to produce the "data_4.pkl" file from the "Data-Creation.ipynb" notebook and put it in the same folder as the notebook.
Notebook containig the visualization of the genres. To run it you need to produce the "data.pkl" file from the "full_data_creation.ipynb" notebook and put it in the same folder as the notebook.
Notebook extracting the whole data from the dataset and creating the "data.pkl" file.
Notebook extracting the data from the dataset only for the 4 genres: "Rock","Hip-hop","Experimental" and "Electronic" and creating the "data_4.pkl" file.