Classification of Turkish makam music: a topological approach

Abstract

In this paper, we study Turkish makam music, a system of varied melodies and chords, computationally. Our main goal is to classify the makams using their notes. For this reason, we utilize the topology of complex networks. We first represent songs with weighted networks where vertices and edges correspond to musical notes and their co-occurrences respectively. We then define the diffusion Fréchet function over the weighted networks to encode the network topology and finally reach our goal by combining the function values with machine-learning algorithms. Our experiments show that such network representation with the diffusion Fréchet function is promising in classifying makam music and more effective than the n-gram technique, which is the most-used automated makam classification method. We believe that our method can be extended to any music, not only makam music.