High-dimensional covariates often admit linear factor structure. To effectively screen correlated covariates in high-dimension, we propose a conditional variable screening test based on non-parametric regression using neural networks due to their representation power. We ask the question whether...

Networks contain complex patterns of dependency and require multiple levels of analysis to explain their formation, structure, and outcomes. In this Element, the authors develop the Multilevel Network Framework. The framework serves as (i) a conceptual tool to think more deeply about network...