This is a project that classify 4 genre of music(jazz,pop,classical,rock), The repository consist of a classification program, the GTZAN dataset and pyAudioAnalysis library.
The program first extract 4 features named ZCR, energyEntropy, SpectralEntropy, Spectral Flux, 13-D MFCC. Then, we get the mean-feature-vector from each music. After that, We train the classifer by 50 tracks from each genre. Finally, we yield the result by predicting labels for other 25 songs.
###Run the code
- Download python,pip and the repository (Try not to run python on windows, may got some solvable problems)
- Goto
genre
directory pip install -r requirement.txt
- run on bash:
python m_classify.py
run on python shell:python
thenexecfile('m_classify.py')
- Alternatively,you can run it on The Jupyter Notebook (a matlab-alike interactive environment)
###Unfinished task
The task of data mining should use cross-validation set instead of only training and testing set. We can use the library of sklearn for improving performence measure.
Also, find other ways to have better accuracy.
####more info:
http://scikit-learn.org/stable/modules/cross_validation.html
http://mirlab.org/jang/books/audioSignalProcessing/appNote/musicGenreClassification/html/goTutorial.html