Home  /  Ricerca  /  Complex networks and deep learning for copper flow across countries

Complex networks and deep learning for copper flow across countries

In this article, we utilize complex network models to represent the extraction, refining, and processing chain of copper on a global scale between 2017 and 2021. For each year, we examine the import-export relationship network of raw copper minerals, processed copper, waste products, and machinery for the metallurgical industry. By approximating the trade networks with multilayer blockmodel models, we annually identify 5 groups of nations based on their roles in the copper supply chain and reconstruct the major annual changes in global trade. Finally, using a deep learning algorithm, we obtain a representation of each nation's position in the global copper trade, identifying the existence of a densely connected core of nations at the center of international trade.

Read the full paper

By Lorenzo Federico, Ayoub Mounim, Pierpaolo D’Urso, and Livia De Giovanni