基于构造U统计量提出了检验对称随机变量分布对称中心的一种新的检验统计量,获得了该检验统计量在原假设与备择假设下的极限分布,并用Bootstrap方法来获得检验拒绝域的临界值,对几类重要对称分布下该检验统计量的检验功效进行了比较,说明了该检验统计量的优良性.
A new test based on U statistics for the center of symmetric distribution is proposed in this paper. Asymptotic null distribution and asymptotic non-null distribution of this statistic are derived, so that oritical value of rejection region could be obtained by using bootstrap method. The power functions comparision of our test with counting statistic means that our U statistic has good performance for several symmetric distribution.