Bootstraps for Time Series

摘要: We review and compare block, sieve local bootstraps for time series thereby illuminate theoretical aspects of the procedures as well as their performance on finite-sample data. Our view is selective with the intention providing a new fair picture some particular of bootstrapping time series. The generality block bootstrap contrasted sieve bootstraps. discuss implementational advantages disadvantages. argue that two types often outperform method, each them in its own important niche, namely linear categorical processes. Local bootstraps, designed nonparametric smoothing problems, are easy to use implement but exhibit cases low performance.