PALMS: Parallel Adaptive Lasso with Multi-directional Signals for Latent Networks Reconstruction

Networks are commonly existing in our world, which characterize the interactions between different items in many fields, such as the social networks between people and the trading networks between companies. With the deepening of research into various complex dynamic systems, network data and network-based dynamic processes have increasingly become focal points of academic inquiry. From the perspective of empirical analysis, the network structures commonly influence the changes and evolution of the world profoundly (Dhar et al., 2014), like transportation networks between cities, supply chain networks for international trades, and competition relationships in an evolutionary ultimatum game. Many scholars regard the known network structures as a treatment, focusing on whether existing network connections exert an influence on other variables, which is commonly referred to as network effects identification. Examples include the impact of transportation networks on economic development (Bramoullé et al., 2009) and the influence of social networks on U.S. election outcomes (Herzog, 2021; Kleinnijenhuis and De Nooy, 2013).