Martingale approximation for common factor representation
Victor Bystrov, Antonietta di Salvatore · Munich Personal RePEc Archive (Ludwig Maximilian University of Munich) · 2012
In this paper a martingale approximation is used to derive the limiting distribution of simple positive eigenvalues of the sample covariance matrix for a stationary linear process. The derived distribution can be used to study stability of the common factor representation based on the principal component analysis of the covariance matrix.