Causal Network Inference via Group Sparse Regularization

Abstract

This paper addresses the problem of inferring sparse causal networks modeled by multivariate auto-regressive (MAR) processes. Conditions are derived under which the Group Lasso (gLasso) procedure consistently estimates sparse network structure. The key condition involves a "false connection score." In particular, we show that consistent recovery is possible even when the number of observations of the network is far less than the number of parameters describing the network, provided that the false connection score is less than one. The false connection score is also demonstrated to be a useful metric of recovery in non-asymptotic regimes. The conditions suggest a modified gLasso procedure which tends to improve the false connection score and reduce the chances of reversing the direction of causal influence. Computational experiments and a real network based electrocorticogram (ECoG) simulation study demonstrate the effectiveness of the approach.

References in corpus (1)

• Learning Networks of Stochastic Differential Equations

Cited by in corpus (32)

• Learning graphs from data: A signal representation perspective
• Connecting the Dots: Identifying Network Structure via Graph Signal Processing
• Signal Processing on Graphs: Causal Modeling of Unstructured Data
• Network inference and biological dynamics
• Validity of time reversal for testing Granger causality
• Learning the Conditional Independence Structure of Stationary Time Series: A Multitask Learning Approach
• Blind identification of stochastic block models from dynamical observations
• Localizing Changes in High-Dimensional Vector Autoregressive Processes
• SILVar: Single Index Latent Variable Models
• Conditioning of Random Block Subdictionaries with Applications to Block-Sparse Recovery and Regression
• Latent Variable Time-varying Network Inference
• A scalable multi-step least squares method for network identification with unknown disturbance topology
• Joint Association Graph Screening and Decomposition for Large-scale Linear Dynamical Systems
• High Dimensional and Banded Vector Autoregressions
• Graph Learning for Spatiotemporal Signals with Long- and Short-Term Characterization
• Economy Statistical Recurrent Units For Inferring Nonlinear Granger Causality
• Compressed Sensing for Network Reconstruction
• Causal Dependence Tree Approximations of Joint Distributions for Multiple Random Processes
• Bounded Degree Approximations of Stochastic Networks
• Signal Processing on Directed Graphs
• Learning hidden influences in large-scale dynamical social networks: A data-driven sparsity-based approach
• On the Sample Complexity of Graphical Model Selection for Non-Stationary Processes
• From dynamics to links: a sparse reconstruction of the topology of a neural network
• Causal Link Discovery with Unequal Edge Error Tolerance
• Bayesian topology identification of linear dynamic networks
• Online Graph Topology Learning from Matrix-valued Time Series
• Dynamic Networks with Multi-scale Temporal Structure
• Single Index Latent Variable Models for Network Topology Inference
• Dynamic network identification from non-stationary vector autoregressive time series
• Learning sparse linear dynamic networks in a hyper-parameter free setting
• Graphical LASSO Based Model Selection for Time Series
• Learning conditional independence structure for high-dimensional uncorrelated vector processes