This is the final project of Machine Learning course of University of Tehran.
Install ffmpeg on Debian/Linux using bellow command:
apt-get install ffmpeg
Then you have to install any library required based on the file you want to use.
About 270 mp3 file were gathered for this project. Then we splitted files to 30 seconds frames in wav format.
After having files with equal length, we extracted features using librosa with different feature extraction methods like MFCC and Zero Crossing Rate.
Using this, our data was model readable data.
-
Classification
- SVM
- MLP
- KNN
-
Clustring
- DBSCAN
- KMeans
- K Medoids
