Wavelet-Based Estimation Procedures for Seasonal Long-Memory Models

Motivation • Seasonal long memory is gaining attention as a time series model in econometrics and the physical sciences. { Arteche and Robinson (2000) proposed log-periodogram regression and Gaussian semi-parametric estimation for seasonal long-memory processes. • The discrete wavelet transform (DWT) is a useful alternative to the discrete Fourier transform for the analysis and synthesis of long-memory (LM) processes. • Can the success of wavelet techniques for LM processes be duplicated here? • Must generalize the band-pass filtering structure of the DWT. ⇒ the discrete wavelet packet transform (DWPT).