Using the Bootstrap to Test for Symmetry Under Unknown Dependence
提出用自回归筛子自助法近似检验统计量的零分布，以检验分数积分过程单变量边际分布的对称性，无需知道记忆参数或渐近分布，并通过蒙特卡洛实验和实际数据验证。
This article considers tests for symmetry of the one-dimensional marginal distribution of fractionally integrated processes. The tests are implemented by using an autoregressive sieve bootstrap approximation to the null sampling distribution of the relevant test statistics. The sieve bootstrap allows inference on symmetry to be carried out without knowledge of either the memory parameter of the data or of the appropriate norming factor for the test statistic and its asymptotic distribution. The small-sample properties of the proposed method are examined by means of Monte Carlo experiments, and applications to real-world data are also presented.