Predictive modeling of dynamic cluster structure of popular music creation styles

Abstract

**Author(s):** Nakamura, E; Singh, R While previous studies on music evolution focused on analysis of trends in features in historical data, construction of a mathematical model that can predict the evolution of the distribution of musical products remains as a challenge. This is especially because the distribution of music styles in a society has a complex structure consisting of clusters that dynamically change. To understand these complex dynamics, we developed statistical modeling methods and analyzed two large datasets of popular music created in Japan and the US. We found that intra-cluster dynamics, such as the contraction and shift of a cluster, as well as inter-cluster dynamics represented by clusters' relative frequencies, often exhibit notable dynamical modes that hold across different cultures. Moreover, we developed an evolutionary model incorporating these dynamical modes and showed that it is effective for predicting the future creation style distributions. Our results indicate the importance of an accurate model of cluster dynamics for predicting future distributions of cultural artefacts. The proposed methodology is general and can potentially be applied to other domains of creative culture.