Data synchronization via node degree in a network of coupled phase oscillators

As a new method for data synchronization, a modified version of the Kuramoto model is introduced to solve the clustering problems. The modified model can generate data clusters with various statistical disributions by setting the natural frequencies of the oscillators to the node degrees of a complex network conveying information on multivariate data. Unlike the original method of data synchronization, the proposed method allows us to deal with a dataset having clusters with non-convex shapes. Through three case studies, we show that the proposed method outperforms existing data-clustering algorithms, such as the Density-Based Spatial Clustering of Applications with Noise.