Complex network approach to classifying classical piano compositions

Department of Physics and State Key Laboratory of Surface Physics, Fudan University - Shanghai 200433, China

^(a) jphuang@fudan.edu.cn

Received: 5 August 2016
Accepted: 29 October 2016

Abstract

Complex network has been regarded as a useful tool handling systems with vague interactions. Hence, numerous applications have arised. In this paper we construct complex networks for 770 classical piano compositions of Mozart, Beethoven and Chopin based on musical note pitches and lengths. We find prominent distinctions among network edges of different composers. Some stylized facts can be explained by such parameters of network structures and topologies. Further, we propose two classification methods for music styles and genres according to the discovered distinctions. These methods are easy to implement and the results are sound. This work suggests that complex network could be a decent way to analyze the characteristics of musical notes, since it could provide a deep view into understanding of the relationships among notes in musical compositions and evidence for classification of different composers, styles and genres of music.