Generating fine-grained surrogate temporal networks

Abstract Temporal networks are essential for modeling and understanding time-dependent systems, from social interactions to biological systems. However, real-world data to construct meaningful temporal networks are expensive to collect or unshareable due to privacy concerns. Generating arbitrarily large and anonymized synthetic graphs with the properties of real-world networks, namely surrogate networks, is a potential way to bypass the problem. However, it is not easy to build surrogate temporal networks which do not lack information on the temporal and/or topological properties of the input network and their correlations. Here, we propose a simple and efficient method that decomposes the input network into star-like structures evolving in time, used in turn to generate a surrogate temporal network. The model is compared with state-of-the-art models in terms of similarity of the generated networks with the original ones, showing its effectiveness and its efficiency in terms of execution time. The simplicity of the algorithm makes it interpretable, extendable and scalable.