B1754
Title: A simulation and estimation algorithm for residual correlation analysis of long-range dependence (LRD) FANOVA models
Authors: Diana Paola Ovalle-Munoz - University of Granada (Spain) [presenting]
Maria Dolores Ruiz-Medina - University of Granada (Spain)
Abstract: A simulation algorithm is derived for the generation of multifractional time series models in the context of infinite-dimensional processes. In particular, multifractionally integrated functional autoregressive moving average processes are generated, displaying different orders of long-range dependence (LRD) in time at each spatial resolution level. Applying previous results by Ruiz-Medina on weak-consistent minimum contrast parameter estimation based on the periodogram operator, an estimation algorithm is implemented to approximate the long-memory operator, characterizing the strong dependence structure of these models. The application of these algorithms in the implementation of residual correlation analysis in the context of FANOVA models with multifractionally integrated functional autoregressive moving average noise is illustrated as well.