On confidence intervals and tests for autocorrelations (Q1083819): Difference between revisions

We propose a statistical procedure for estimating the asymptotic variances and covariances of sample autocorrelations from a stationary time series so that confidence regions and tests on a finite subset of autocorrelations can be implemented. The corresponding algorithm is described. The accuracy of the asymptotic confidence intervals for finite samples is studied by Monte Carlo simulations. Further, our method is illustrated with examples from the literature.