Dynamic Networks with Multi-scale Temporal Structure

We describe a novel method for modeling non-stationary multivariate time series, with time-varying conditional dependencies represented through dynamic networks. Our proposed approach combines traditional multi-scale and network based neighborhood selection, aiming at capturing temporally local structure in the data while maintaining sparsity of potential interactions. framework is on recursive dyadic partitioning, which recursively partitions temporal axis into finer intervals allows us to detect structural changes varying resolutions. The selection achieved penalized likelihood estimation, where penalty seeks limit number neighbors used model data. present theoretical numerical results describing performance our method, motivated illustrated using task-based magnetoencephalography (MEG) neuroscience.