Time-varying Group Lasso Granger Causality Graph for High Dimensional Dynamic system
作者：
Highlights：
• To identify the structure change characteristics of causal relationships for time-varying networks, we propose dynamic network based on Granger causality for modeling the time-varying directed dependency structures.
• For the structural learning problem of the proposed time-varying Granger causality graph, we introduce a kernel reweighted group lasso method. The method considers the group structure of the lagged variables and improves the accuracy and efficiency of the algorithm.
• In addition, the time-varying Granger causality network is applied to financial field. The results show that networks based on Granger causality have rich indicators to characterize both the global evolution features of networks and the different functions of individual nodes in the graph.
摘要
•To identify the structure change characteristics of causal relationships for time-varying networks, we propose dynamic network based on Granger causality for modeling the time-varying directed dependency structures.•For the structural learning problem of the proposed time-varying Granger causality graph, we introduce a kernel reweighted group lasso method. The method considers the group structure of the lagged variables and improves the accuracy and efficiency of the algorithm.•In addition, the time-varying Granger causality network is applied to financial field. The results show that networks based on Granger causality have rich indicators to characterize both the global evolution features of networks and the different functions of individual nodes in the graph.
论文关键词：Time-varying Granger causality,Feature selection,Group Lasso,Financial market network
论文评审过程：Received 11 May 2020, Revised 13 April 2022, Accepted 11 May 2022, Available online 20 May 2022, Version of Record 26 May 2022.
论文官网地址：https://doi.org/10.1016/j.patcog.2022.108789
