HCGA: Highly comparative graph analysis for network phenotyping

Robert L. Peach, Alexis Arnaudon, Julia A. Schmidt, Henry A. Palasciano, Nathan R. Bernier, Kim E. Jelfs, Sophia N. Yaliraki, Mauricio Barahona

Abstract: The wide applicability of graphs to many scientific domains has fueled a large corpus of network theoretical literature. Here, we introduce HCGA, a package that computes a comprehensive collection of network-based properties and uses them as features for the highly comparative analysis of graph datasets. Our methodology achieves state-of-the-art accuracy on benchmarks and provides interpretable insights into novel datasets.