A0635
Title: Eigen-analysis for functional time series
Authors: Yanrong Yang - The Australian National University (Australia) [presenting]
Yuan Gao - The Australian National University (Australia)
Han Lin Shang - Australian National University (Australia)
Yang Yang - University of Newcastle (Australia)
Abstract: Spectral analysis is important in dimension reduction for functional data. Under a general data structure for functional time series, the influence of temporal dependence is studied on empirical eigenvalues and eigenvectors from the sample covariance operator. Asymptotic properties of empirical eigenvalues are established to quantify such influence. Based on the developed theory, a new algorithm is proposed to recover the non-stationary subspace and the principal component subspace. Various simulations are constructed and empirical analysis includes eigen-analysis on global mean temperature data and Australian mortality data.
