基于秩的持久性变化的监测

在滑动残差比率检验统计量的基础,上基于秩提出了新的非参检验方法,并在原假设下得到其渐近分布,在备择假设下证明其容许性.模拟结果表明利用非参方法对厚尾情形所犯第一类错误更小.最后将方法应用到美国密歇根消费者信心指数初值,结果表明该方法是有效的.

Monitoring the changes in persistence based on ranks

In this paper,a new nonparametric test is constructed to monitor the change in persistence on the ranks of sequences,and the asymptotic distribution under the null hypothesis is investigated. The consistence of the test is shown under the alternative hypothesis. The Monte Carlo simulations demonstrate that the proposed test is more accurate when the innovations are heavy-tailed. An application of Michigan's index of consumer confidence shows the validity of the proposed test.