Title: A bootstrap test for sphericity of time series
Authors: Yan Liu - Kyoto University; RIKEN AIP (Japan) [presenting]
Abstract: The problem of testing the sphericity hypothesis for the covariance matrix of high-dimensional time series is considered. It has been shown that the test statistic for sphericity under the null hypothesis is asymptotically normal for high-dimensional time series models. However, the asymptotic variances of the test statistic are different from each other if the time series models are different. To alleviate the computational complexity and model dependence, we propose a bootstrap procedure for the test statistic. The validity and the consistency of the proposed method are justified from the asymptotic theory. The performance is illustrated by simulation studies.