|
Abstract:
|
Vector Autoregression (VAR) models are popular in modeling multivariate time series in different fields such as econometrics, engineering etc. However, these models are computationally complex with the number of parameters scaling quadratically with the number of time series. We propose a sparse modeling approach called Neighborhood Vector Autoregression (NVAR) to efficiently analyze high-dimensional multivariate time series. We assume that the time series have underlying distances among them based on the inherent setting of the problem, e.g., due to spatial dependency. When this distance matrix is known (or can be calculated directly), we demonstrate that our NVAR method provides a consistent estimation of the parameters. The performance of the proposed method is compared with other existing approaches in both simulation studies and a real-world application.
|
