Identification of Canonical Models for Vectors of Time Series: A Subspace Approach

Abstract

We propose a new method to specify linear models for vectors of time series with some convenient properties. First, it provides a unified modeling approach for single and multiple time series, as the same decisions are required in both cases. Second, it is scalable, meaning that it provides a quick preliminary model, which can be refined in subsequent modeling phases if required. Third, it is optionally automatic, because the specification depends on a few key parameters which can be determined either automatically or by human decision. And last, it is parsimonious, as it allows one to choose and impose a canonical structure by a novel estimation procedure. Several examples with simulated and real data illustrate its application in practice.