A Simulation Comparison of Bootstrap Procedures in Periodically Correlated Time Series Models

Abstract

The presence of periodicity in data with periodic structure has become an important issue in parameter estimation. Several methods have been studied with intention estimating different parameters or constructing confidence intervals for the parameters. In this paper we investigate the performance of the bootstrap procedures designed for dependent data in the case of Periodically Correlated time series models. Several models with periodic structure are studied in this paper and we use R programming language to realize a simulation comparison of the performance of bootstrap procedures presented.