Stationary subspace analysis of nonstationary covariance processes: eigenstructure description and testing (Q2214252)

The following is a summary of the abstract. The authors perform Stationary Subspace Analysis (SSA) for zero mean nonstationary covariance processes. SSA searches for linear combinations of the components of nonstationary vector time series that are stationary. The stationary subspaces and their dimensions are characterized in terms of eigenvalues and eigenvectors of certain symmetric matrices. This characterization is used to derive formal statistical tests for estimating dimensions of stationary subspaces. Eigenstructure-based techniques are also proposed to estimate stationary subspaces, without relying on previously used computationally intensive optimization-based methods. The introduced methodologies are examined on simulated and real data.