Althoitgh music classification has been widely discussed, relatively few stitdies have explored the topic of polyphonic music classification. On the other hand, Bellini et al., 2005 revealed the importance of integrating symbolic music representation with mitltimedia objects due to the development of the MPEG standard. Therefore, it is foreseen that more and more music objects in symbolic format and mitltimedia objects, integrated with symbolic music representation, will be published via the Internet, as video and audio presentations. To cater to future music-related applications of these symbolic mv representations, we propose a novel approach to music classification by discovering high-level featitres of polyphonic music from the score. Overall, the main goal of this paper is to present a new conceptital framework that can automatically parse MusicXML files and extract their qualitative and quantitative features. In addition, we also propose appropriate featitres to improve classification accitracy and create an effective classifier for automatic mitsic classification. To assess the proposed approach, music features extracted from a score were used to test the music classification accunracy by employing the decision tree-based (C4.5), SVM, and k-NN methods. The experimental results show that the proposed approach substantially outperforms other methods in terms of classification accuracy.
|Number of pages||11|
|Journal||International Journal of Innovative Computing, Information and Control|
|State||Published - 1 Dec 2009|
- Data mining
- Music classification
- Music feature
- Music representation