Classification of Turkish Maqam music using k-means algorithm and artificial neural networks

In this study a proposal is made to classify the maqams of songs in Turkish Maqam Music. The MIRToolbox (a toolbox for MATLAB) is used to extract the features which are used in audio and music processing. Obtained data are preprocessed by principal component analysis. The songs are clustered by k-means algorithm and classified by feed forward artificial neural networks according to their maqams.